Soft synthesis and characterization of goethite-based nanocomposites as promising cyclooctene oxidation catalysts

Goethite based nanocomposites with a different composition such as 6FeO(OH)·MnO(OH)·0.5H2O (Mn-composite), xFeO(OH)·M(OH)2·yH2O (Co-composite (M: Co, x = 12, y = 3), Ni-composite (M: Ni, x = 7, y = 2)) and xFeO(OH)·MO·yH2O (Cu-composite (M: Cu, x = 5.5, y = 3), Zn-composite (M: Zn, x = 6, y = 1.5)) have been prepared by a soft chemical synthesis consisting in acetate hydrolysis. The data provided by Fourier transform infrared (FTIR), ultraviolet-visible-near infrared (UV-Vis-NIR), electron paramagnetic resonance (EPR) and Mössbauer spectra account for a slight modification of all composites' physicochemical properties compared to the starting material. Powder X-ray diffraction and transmission electron microscopy (TEM) investigations revealed the secondary phase nature and presence along with that of goethite. The TEM data are also consistent with a nano rod-like morphology with a 5–10 nm width and an average length of 40 nm. The catalytic oxidation of cyclooctene with O2 using isobutyraldehyde as reductant and acetonitrile as a solvent was performed in batch conditions for 5 h at room temperature. The selectivity for the epoxide was higher than 99% for all tested solids. The conversion of cyclooctene decreased from 55% to 4% following the same order of variance as the base/acid sites ratio: Mn-composite > Fe-composite > Co-composite > Ni-composite > Zn-composite > Cu-composite. The 6FeO(OH)·MnO(OH)·0.5H2O (Mn-composite) exhibited the most promising catalytic activity in cyclooctene oxidation, which can be correlated with the redox ability of Mn(iii) combined with the increased base character of this solid. The catalytic activity of this sample decreases by 10% after several successive reaction cycles.


Introduction
The low production costs and their chemical stability against UV radiation or corrosive agents make inorganic pigments one of the most important species of this type. Among these, both iron oxides and oxyhydroxides such g-Fe 2 O 3 (maghemite), Fe 3 O 4 (magnetite), a-FeO(OH) (goethite), and FeO(OH)$H 2 O (limonite) are constituents of many paints, enamels, and varnishes. 1,2 On the other hand, all ores enriched in such compounds represent the raw material for the cast or steel iron industry. 3,4 Some concrete based on limonite or ilmenite-limonite is used as nuclear reactor shielding. [5][6][7] Both goethite and limonite are used as precursors for preparing many functional materials such as magnetic oxides, maghemite, magnetite, and ferrites. [8][9][10] Moreover, the limonite addition to lithium-ion batteries increases the ionic resistance and conductivity. 11,12 A limonite-doped lithium borate glass was developed as gammaray shielding material. 13 Natural limonite was also used as raw material to produce a nanosized zero-valent iron by hydrogen reduction with superior performance on p-nitrophenol decomposition compared with a commercial iron powder. 14 Furthermore, a cotton fabric coated with a polymer containing a mixture of goethite, limonite, and hematite as additives exhibited bacteriostatic and antibacterial effects against Staphylococcus aureus and Escherichia coli. 15 Far from these various applications, other studies use hydrated iron oxyhydroxide as a catalyst for inorganic and organic processes. Examples of such methods are arsenite oxidation, 16 arsenic removal from wastewater, 17 ammonia, 18 pyridine 19 or hydrogen sulphide 20 removal from coke oven gas, coal liquefaction for oil production, 21 organic compounds removing from wastewater [22][23][24][25] and microcystin-LR hydrolysis in cancer prevention. 26 On the other hand, iron-containing systems containing bimetallic [27][28][29][30][31][32][33] or trimetallic layered oxyhydroxides 27,[34][35][36] were developed, based on their tunable electronic structures and rich active sites, 37 as valuable nonprecious metal-based materials for oxygen evolution reaction (OER).
Synthesis of this kind of compounds with controlled size, shape, morphology, iron substitution, and the active surface is important. Moreover, since their properties are affected by all these factors, they also need to be precisely tuned depending on the requirements of particular applications. 38,39 Usually, goethite is mainly obtained by the wet-chemical precipitation process starting from water-soluble iron salts (chloride, nitrate, or sulfate) by adding either caustic soda or ammonia in the presence of air as oxidizing agent. 40 Parameters like pH, salt concentration, temperature, and stirring velocity are involved in the particle size and geometry, which control the properties. 38,39 However, the attention in this eld is now focused on the development of some non-conventional methods such as plasma treatment, 24,25 electrodeposition, 27,30,31,34 chemical deposition, 29,32 or chemical deposition assisted by a magnetic eld. 33 Moreover, all these methods are eco-friendly and control both the nano-dimension of the particles and the substitution degree.
In the present work, we extended the study to obtain manganese, nickel, cobalt, copper, and zinc goethite-based nanocomposites by a so chemical method successfully used for some ferrites synthesis. 41 Furthermore, the inuence of the second metallic ion from these nanocomposites on their catalytic behavior in cyclooctene epoxidation with molecular oxygen, in the presence of isobutyraldehyde, under ambient conditions was studied as well.

Reagents
We used high purity reagents purchased from Sigma-Aldrich (Saint-Louis, MO, USA) ( ammonia 25%, H 2 O 2 ) as received without further purication. Cyclooctene, isobutyraldehyde, and acetonitrile were also purchased from Sigma-Aldrich and have been previously distilled before reactions. Gaseous O 2 with 99.99% purity purchased from Linde gas was used as an oxidant agent.

Instruments and methods
Chemical analysis of iron, manganese, cobalt, nickel, copper, and zinc was performed using the usual micro methods 42 aer sample dissolution with hydrochloric acid. The iron was precipitated with ammonia and gravimetrically determined aer calcination, while the other metallic ions were determined from the resulted solution. The heating curves (TG, DTG, and DTA) were recorded using a Labsys 1200 instrument (Setaram, Caluire, France), with a sample mass of about 20 mg over the temperature range of 293-1173 K, using a heating rate of 10 K min À1 . The measurements were carried out in a synthetic air atmosphere (ow rate of 16.70 cm 3 min À1 ) using alumina crucibles. The Fourier transform infrared (FTIR) spectra were recorded with a Spectrum BX II (Perkin Elmer, USA) spectrometer in the 350-4000 cm À1 range by accumulating 32 scans at a resolution of 4 cm À1 . The powdered samples were diluted into KBr powder in a 1 : 100 mass ratio, ground thoroughly, and pressed into pellets. UV-Vis spectroscopy was performed in solid-state on a V 670 spectrophotometer (Jasco, Easton, MD, USA) with Spectralon as standard in the 200-1500 nm range. The electron paramagnetic resonance (EPR) spectroscopy measurements were carried out with a Bruker EMX premium X (Bruker, Karlsruhe, Germany) equipped with an X-SHQ 4119HS-W1 X-Band resonator at a microwave frequency of 9.4457 GHz and power of 0.06325 mW. Further measurements parameters were: conversion time 10 ms, time constant 5.12 ms, modulation amplitude 0.3 mT with one scan. We used a digital temperature control system ER 4131VT with a liquid nitrogen cryostat from Bruker (Bruker, Karlsruhe, Germany) for cooling. The 57 Fe Mössbauer spectra have been obtained in transmission geometry, at 6 K and room temperature, by inserting the samples in a close cycle Janis cryostat (Edina, Minnesota, USA). A SEECO-type spectrometer (Edina, Minnesota, USA) operating under the constant acceleration mode and a 57 Co(Rh) radioactive source of about 30 mCi activity were used. The acquired 57 Fe Mössbauer spectra were analyzed using the NORMOS soware, which allows the decomposition of the measured absorption pattern in spectral components corresponding to different iron non-equivalent positions. In the case of a continuous distribution of the hyperne parameters, the tting procedure can be performed accordingly using specic routines that provide the envisaged probability distribution function and complementary average hyperne parameters. The isomer shi is reported relative to the isomer shi of metallic Fe at room temperature. Powder X-ray diffraction (XRD) patterns were recorded with a Bruker D8 Advance X-ray diffractometer (Bruker, Karlsruhe, Germany) (Cu anode and Ni lter, l ¼ 1.54184Å) in Bragg-Brentano conguration. We determined the lattice parameters and the average crystallites size by the Rietveld renement method 43 using the MAUD soware. The JEOL 2100 Transmission Electron Microscope (TEM) (Tokyo, Japan), equipped with energy dispersive X-ray (EDS), has been used for transmission electron microscopy investigations. Specimens have been prepared using the standard powder method. until a brown precipitate formed. The solid product was ltered off, washed several times with water, and air-dried.
The syntheses of goethite-nanocomposites were performed by adding 20 mmol freshly prepared goethite to solutions containing 10 mmol M(CH 3 COO) 2 $nH 2 O (M: Mn, Co, Ni, Cu, and Zn) in 100 mL water. Reaction mixtures were heated at 373 K under continuous stirring for 24 h. The obtained brown products were ltered off, washed several times with water, and airdried.

Textural characterization
Nitrogen adsorption-desorption isotherms at 77 K were recorded on a Micromeritics ASAP 2020 automated gas adsorption system (Norcross, GA, USA). The samples were degassed at 90 C for 12 hours under vacuum before analysis. Specic surface areas (S BET ) were calculated according to the Brunauer-Emmett-Teller (BET) equation, using adsorption data in the relative pressure range between 0.05 and 0.30. The total pore volume (V total ) was estimated from the amount adsorbed at the relative pressure of 0.99. The pore size distribution curves were obtained from the adsorption data using a DFT (density functional theory) model.

Catalytic tests
This study evaluates the catalytic activities of the synthesized materials in the oxidation of cyclooctene with molecular oxygen. We used isobutyraldehyde as a reductant and acetonitrile as a solvent in batch conditions. Thus, 20 mg of material were contacted with 0.04 mol cyclooctene and 0.08 mol isobutyraldehyde in 10 mL acetonitrile as a solvent. We performed the catalytic tests for 5 h at room temperature. Molecular oxygen was admitted in a sealed stirred ask of 250 mL by a tube linked to an oxygen pressurized cylinder employing a manometer. During the tests, the oxygen pressure was maintained at 1 atm. Before reactants and the catalyst admission in the reactor, the air was removed by purging molecular oxygen for 5 minutes. During the reaction, at each hour, quantities of 50 mL were extracted from the reactor and analyzed with a Gas Thermo Quest Chromatograph (ThermoFisher Scientic Inc., Waltham, MA, USA) equipped with an FID detector and a capillary column with DB5 stationary phase. All products, as well as reactants, were identied by comparison with standard samples (retention time in GC) and by mass spectrometer-coupled chromatography VARIAN SATURN 2100 T (LabX, Midland, ON, Canada) to evaluate the cyclooctene conversion and the selectivity to epoxide.

Base sites determination
The total number of base sites was determined using the irreversible adsorption of acrylic acid (pK a ¼ 4.2). Samples of dried solids (0.05 g) were contacted for 2 hours (duration required for reaching the equilibrium in the liquid-solid system) with 10 mL of 0.01 M solution of acrylic acid in cyclohexane in brown sealed bottles under mild stirring (150 rpm) at room temperature. It was assumed that the interaction of the solids with atmospheric CO 2 and water was negligible since the samples were exposed to the atmosphere only during weighing. The concentration of the acrylic acid remaining in the solution aer reaching equilibrium was determined by UV-Vis spectrometry at l max ¼ 225 nm using the Jasco V-650 spectrometer (Tokyo, Japan). For each solid sample, 3 parallel determinations were performed, and the obtained results were averaged. The amount of acrylic acid (AA) adsorbed was calculated with the formula: where indexes i and f refer to the initial amount and the nal amount of acrylic acid in the solution, respectively.
The method is inspired by the one used to determine base sites in hydrotalcite-type materials. 44 However, in the case of these solids, we could not use phenol for the separate determination of strong base sites since phenol is known to give colored combinations with iron.
The total concentration of base sites was calculated with the formula: where wt is the weight of the solid sample.

Acid sites determination
The total number of acid sites was determined by pyridine adsorption. The distribution of acid sites, Lewis and Brönsted, respectively, were calculated from the areas of the corresponding peaks in the DRIFT spectra recorded on FT/IR-4700 Jasco spectrometer (Tokyo, Japan). Samples of dried solids (0.05 grams) were contacted with pyridine aliquots (0.2 mL each) and maintained under inert ow at 90 C for the removal of physisorbed pyridine. The procedure was repeated until the weight of the sample aer two consecutive additions of pyridine was constant (did not vary with more than 0.0001 g). Then, the DRIFT spectrum of the sample with adsorbed pyridine was recorded, considering the DRIFT spectrum of the freshly dried solid as background. According to literature data, [45][46][47] the bands corresponding to pyridine adsorbed on Lewis acid sites appear in the ranges of 1435-1455 and 1570-1615 cm À1 while those corresponding to pyridine adsorbed on Brönsted acid sites appear in the range of 1520-1555 and at 1630 cm À1 . In the ESI, † we have included Fig. S1 † with the DRIFT spectra of the pyridine adsorbed on the samples.

Results and discussions
In this paper, we report the synthesis and characterization of some goethite based nanocomposites of type: These samples were characterized as nanocomposites by chemical and thermal analysis, IR, UV-Vis-NIR, EPR, and Mössbauer spectroscopy. Simultaneously, their morphology and particle dimension was provided by powder X-ray diffraction and TEM studies. Textural parameters were calculated from N 2 sorption isotherms.

Chemical analysis and thermal decomposition
We deduced the composition of these species from the metal contents, the mass of water lost up to 423 K (W 1 ), the mass of water lost up to 633 K (W 2 ), and residue formed at 873 K during the thermal decomposition, as reported in Table 1.
The thermal decomposition of all samples occurs in two endothermic steps. The rst step corresponds to crystallization water elimination, which occurs up to 413-523 K, followed immediately by the same compound elimination due to oxyhydroxide/hydroxide decomposition.

IR spectroscopy investigations
Iron oxyhydroxides, FeO(OH)$nH 2 O consists of a vast array of oxo and hydroxo groups with a three-dimensional structure in which the layers are connected through hydrogen bonds. The basic units in these polynuclear species that generate specic bands in the IR spectra of goethite consist of groups of two or three Fe(III) ions links by oxo or hydroxo bridged units. Essential bands noticed in the spectra of synthesized goethite-based nanocomposites are summarized in Table 2. The infrared spectra are presented in Fig. S2 in the ESI. † In the 3130-3410 cm À1 region, one can notice two broad bands assigned to n(OH) asymmetric and symmetric stretching vibration modes for lattice water. The bands around 1600 (n(OH)), 890, 800 (r r (OH 2 )), and 600 cm À1 (r w (OH 2 )) are also due to the water molecule vibration modes. Moreover, the last ones indicate that some of these molecules are coordinated to the metallic centers. The hydroxo groups are distinguished from the aqua ones by the appearance of bands in the range 1115-1140 cm À1 assigned to d(M-OH). Two bands assigned to metaloxygen stretching vibrations appear in all spectra in the 400-460 cm À1 range. 48

UV-Vis-NIR spectroscopy characterization
The UV-Vis-NIR spectroscopy proved to be a valuable tool to establish the oxidation state of transition ions. We used the goethite itself for baseline calibration and as a reference to eliminate the iron interference. In this condition, this method revealed only the d-d bands characteristic for the second

EPR spectroscopy characterization
We recorded each compound's EPR spectra at different temperatures beginning with 150 K being increased in 10 K Fig. 1 Temperature dependency of the EPR spectra for FeO(OH) (a), Co-composite (b) and Zn-composite (c). Temperature dependency of the peak-to-peak line width of the EPR spectra of the goethite and goethite-based nanocomposites (d). The temperature was varied in 10 K per step in the range 150-300 K.  steps until 300 K. The obtained characteristic spectra are presented in Fig. 1(a-c) for goethite, Co-composite, and Zncomposite samples. The EPR spectra for the other goethitebased nanocomposites are shown in the ESI (Fig. S3 †).
The very broad EPR signal observed for all samples highlights the FeO(OH) strong ferromagnetic properties at low temperature. By increasing the temperature, the ferromagnetic properties vanish slowly, which is pointed out in Fig. 1d. The peak-to-peak line width of the EPR signal is plotted against the temperature. Aer reaching the Curie temperature, the material would become paramagnetic, and the peak-to-peak line width would present no changes. Since the Curie temperature of a-FeO(OH) is located at about 900 K, 50 this point was not reached in this study.
Regarding the changes in the peak-to-peak linewidths of the EPR spectra as a function of the temperature, it is clear that the used elements inuenced the ferromagnetic behavior of the FeO(OH). Fig. 1d summarizes these changes, where cobalt induces the most signicant changes in ferromagnetic behavior. The other elements presented in the composite network cause small but visible changes. Thus, the peak-to-peak line width temperature dependency is similar to the FeO(OH)$ H 2 O but different. This indicates a shi in the Currie temperature of the composites.

Mössbauer spectroscopy characterization
Due to the preparation conditions, a substitution effect of goethite can be considered. Therefore, we recorded the Mössbauer spectra of both goethite and goethite-based nanocomposites have to conrm or exclude this effect. Mössbauer absorption spectra obtained at 6 K on FeO(OH) and nanocomposite samples are shown in Fig. 2. All the spectra show a sextet pattern characteristic to a-FeO(OH) at low temperatures as the main component. The    The room temperature experiments provide only ESI, † and the results are shown in Fig. 3.
The effects related to the presence of Mn(III), Co(II), Ni(II), Cu(II), and Zn(II) in the a-FeO(OH) structure have been followed according to the variation of the hyperne parameters. The t results are shown in Table 3. The quadrupole splitting (QUA) and the magnetic hyperne eld BHF values are in good agreement with specic values of a-FeO(OH). The isomer shi (ISO), as well as the spectral line width (WID), do not show signicant variations among the investigated samples.
In the case of the Mn-composite, the additional presence of a quadrupole doublet can be associated with a small fraction of very ne FeO(OH) nanoparticles in a magnetic relaxation regime. According to this specic evolution of the Mossbauer hyperne parameters (almost unchanged among the analyzed samples), we conclude that the structure of the a-FeO(OH) remains unaffected by the presence of the second transition metallic ion.

X-ray powder diffraction
The main crystalline phase was identied for all samples as a-FeO(OH) (PDF. No. 1008766). Fig. 4 shows the X-ray powder diffractograms for goethite and goethite-based nanocomposites.
From the Rietveld renement data (Table 4), it can be observed that the mean crystallite size varies slightly between the goethite and the goethite-based nanocomposites. Such an effect may appear if a doping process of the goethite is considered. Moreover, except for Zn-composite, a slight increase in the lattice parameters can be observed for the other goethite-based nanocomposites. The Zn-composite lattice parameters are approximately the same as for FeO(OH)$H 2 O. As shown in Fig. 4 and Table 4, for the Cu-composite and Zncomposite, along with the a-FeO(OH) phase, another phase(s) appears. In the X-ray diffractogram of Cu-composite, a CuO phase (tenorite) was identied based on the JCPDS no. 1526990, with the lattice parameters determined by Rietveld renement analysis as presented in Table 4. Based on the two intense diffraction peaks at small angles (12 ) and the other three peaks at 24.2, 24.8, and 27.3 , copper(II) acetate (JCPDS no. 00-027-1126) can be identied along with goethite and copper(II) oxide. The X-ray diffractogram of Zn-composite contains diffraction peaks assigned to the ZnO phase with a hexagonal structure based on the JCPDS no. 89-1397.

Transmission electron microscopy investigations
In Fig. 5, the TEM image and corresponding diffraction pattern acquired on FeO(OH) sample, as a reference, show a nano rodlike morphology and the a-phase of FeO(OH) (CIF 1008766). The crystalline nanorods have a relatively uniform width (5-10 nm) and a large distribution of lengths, with an average of roughly 40 nm.
The FeO(OH) phase's morphology tends to be preserved along with the entire series of samples, as shown in Fig. 6. It is worth mentioning that the $20 nm size as provided by Rietveld analysis for the FeO(OH) phase in all cases is a very rough estimation, given the elongated morphology of the entities. The energy-dispersive X-ray spectroscopy (EDX) maps in Fig. 6 shows a specic aggregation of the doping element for all the nanomaterials. In the case of Cu-composite, the TEM investigations correlated with Rietveld renement suggest the formation of a CuO (JCPDS no. 526990, tenorite) secondary phase. In the case of the Zn-composite, the elemental mapping (Fig. 6), the TEM image, and electron diffraction (Fig. 7, a and b) show the presence of a relatively thin layer of highly textured ZnO along with the familiar FeO(OH) phase. This solves the issue regarding the unidentied secondary phase from the Rietveld renement mentioned above.
For the rest of the samples (Mn-composite, Ni-composite, and Co-composite), the elemental mappings analyzed in correlation with the powder X-ray diffraction results suggest the formation of incoherent Mn(III), Ni(II), respectively Co(II) rich entities.

Textural characterization
Textural characterization of the samples was carried out by N 2 adsorption-desorption analysis. According to the IUPAC clas-sication, all the isotherms (Fig. 8) are of type IV. 51 It can be noticed that in the region of low relative pressures (p/p 0 ) up to 0.4, the amount of N 2 adsorbed increases sharply. As p/p 0 rises above 0.4, the uptake of N 2 is slower, while at p/p 0 values higher than 0.8, the adsorption curve tends to atten. All isotherms display H2 hysteresis loops due to capillary condensation in mesopores, whose area differs depending on the pore size. The pore size distribution (insets of Fig. 8) obtained using a DFT model is multimodal. The average pore size is around 3 nm for goethite and M-composites (M: Mn, Co, Ni and Zn), while for the Cu-composite, it is 4.05 nm. The goethite has a specic surface area (S BET ) of 394 m 2 g À1 , while for the composites, the S BET is lower but over 300 m 2 g À1 ( Table 5). The obtained S BET values are much higher than those reported by other authors for similar materials. 52

Catalytic activity
In addition to the large numbers of chemical reactions that show a peculiar scientic and economic interest, oxidation plays a crucial role. Selective oxidation, in particular selective oxidation of alkenes, is by far the most widely used method by which oxygen atoms can be inserted into molecules without leading to mineralization. 53 Epoxides are important functional intermediates playing a crucial role in pharmaceuticals, pesticides, cosmetics, and materials production. 54 However, the selective oxidation of alkenes requires complexes based on homogeneous [55][56][57] or heterogeneous [58][59][60][61] catalysts. Almost all of these catalysts have been used in the presence of nonenvironmentally friendly oxidizing agents such as NaClO, 62 PhIO, 63 dioxiranes, 64 tert-butylhydroperoxide, 65 or potassium peroxomonosulfate. 66 Hydrogen peroxide or molecular oxygen are viable alternatives to those described above. Up to date, there are several studies concerning epoxidation of alkenes in aerobic conditions using transitional metal complexes/support as catalysts and isobutyraldehyde as reductant, which are carried out in mild conditions. 67,68 Total conversion of cyclooctene to the corresponding epoxide was recently reported for catalysts consisting of Co(II) complexes with Schiff base ligands supported on silica-coated magnetite tested under reux conditions. 69 However, the synthesis of such catalysts is difficult and expensive, and the energy consumption for heating the reaction mixture is high. Another expensive catalyst that allowed reaching 85% conversion of cyclooctene with 92% selectivity for epoxide at a lower temperature (e.g., 35 C) is a Co(II) coordinated metal-organic framework. 70 Co 3 O 4 nanoparticles encapsulated in the inside  wall of a meso-SiO 2 shell, which allowed reaching 90% conversion at 40 C, 71 manganese doped cerium oxide catalysts, which allowed getting 80% conversion at 100 C, 72 were recently reported as highly active heterogeneous catalysts for cyclooctene epoxidation using isobutyraldehyde and molecular oxygen. The disadvantages of the latter two systems are that the synthesis of Co 3 O 4 encapsulated in meso-SiO 2 shell catalyst is difficult to control. In contrast, in the case of Mn-doped cerium oxide catalysts, the reaction conditions involve high energy consumption. Until now, iron oxide hydroxide (FeOOH) was used as an efficient catalyst only for alcohol oxidation, organic sulde oxidation, and epoxidation of alkenes in the presence of H 2 O 2 73 while the catalytic activity of FeOOH based nanocomposites containing MnO(OH), Co(OH) 2 , Ni(OH) 2 , CuO or ZnO in the epoxidation of cyclooctene with molecular oxygen has not been reported. Such catalytic systems could present interest since their synthesis is not difficult and does not imply high costs. The reaction mechanism of cyclooctene oxidation with molecular oxygen in the presence of isobutyraldehyde is analogous to that suggested by Nam and co-workers. 74 It involves several steps which conrm that the generation of isobutyric acid and the epoxide formation is in a linear dependence (Scheme 1). However, for all the investigated samples, the epoxide selectivity is more than 99%, conrming the unique character of this selective oxidation.
However, several factors limit or improve catalytic activities due to the presence of three various trends. They depend on the cations in the composition of the solids. The lowest cyclooctene conversions of up to 10%, where a plateau level was quickly reached, were obtained with the Zn-composite and Cucomposite samples (Fig. 9). Both Zn(II) and Cu(II) cations do not display a large domain of oxidation states, allowing the cations' oxidation-reduction abilities. This characteristic is in agreement with Nam's conclusions, 74 claiming that the oxidation of olens occurs in the presence of high-valent oxometal intermediates produced by the reaction of the peroxyacid with the metallic ions from the catalyst. Ni-composite and Cocomposite show a similar variation trend while the plateau level of the conversion is reached aer 4 hours. However, the Co-containing catalyst displays better activities than the Ni-containing one due to its better ability to play the oxidationreduction role.
The goethite (FeO(OH)$H 2 O) and Mn-composite show higher activity values than the other samples. The linear dependence of conversion values versus the reaction time is noticeable. Even aer 5 hours, the appearance of a conversion level plateau is not reached.
Considering the result of the acid and base sites determinations presented in Table 6, there is a clear correlation between the ratio of base sites/acid sites and the conversion values obtained with different catalysts aer 5 hours. The manganese presence leads to the best conversion values, which can also be related to the fact that it induces an increased base character of the solid (base sites/acid sites ratio ¼ 4.44).
Moreover, according to the mechanism presented in Scheme 1, the rst step in cyclooctene transformation consists of metal ion reduction. Taking into account the ionization potential (E) associated with each cation presented in the composites (E(Mn 2+ /Mn 3+ ), 33.69 eV; E(Co + /Co 2+ ), 17.03 eV; E(Ni + /Ni 2+ ), 18.03 eV; E(Cu + /Cu 2+ ), 20.29 eV and E(Zn + /Zn 2+ ), 17.89 eV), 75 it is evident that the reverse transformation will involve the same energy but with an opposite sign. This means that the Mncomposite will provide for the rst oxidation step higher energy than the other ones.
Hence, the synergistic effect of the ability to perform oxidation-reduction cycles and the increased base character induced by the involved modifying cations leads to better materials for selective oxidation of olens. The small fraction of very ne FeOOH nanoparticles detected by Mössbauer spectra in the Mn-composite sample could also account for its improved activity. The results obtained with the Mn-composite at room temperature were better than those reported for cyclooctene epoxidation using other simple catalytic systems such as Ni/nanoporous carbon (e.g., 50% conversion). 68 The recycling studies for the best material, i.e., MnFe 2 O x , indicated its reasonable stability since, aer three reaction cycles, the conversion decreases only by 10% (Fig. 10). The decrease of activity from cycle to cycle could be due to the adsorption of the isobutyric acid generated as a by-product on the base sites of the catalyst since the catalyst was only dried and not washed with solvent before being reused.
Scheme 1 The reaction mechanism for cyclooctene oxidation with O 2 in the presence of isobutyraldehyde.

Conclusions
New goethite-based composites containing Mn(III), Co(II), Ni(II), Cu(II), and Zn(II) were prepared and characterized from a physicochemical and morphological point of view. The Mössbauer spectra evidenced only the goethite pattern even at low temperatures. In contrast, the different ferromagnetic behavior evidenced by EPR spectra and the characteristic d-d bands is consistent with the second transition ion presence in the composite network. The lattice parameters vary only slightly from a composite to other. Powder X-ray diffraction and TEM data indicate a CuO and ZnO phase in the corresponding composite. On the other hand, the elemental mappings suggest the rich entities based on Mn(III), Ni(II), and Co(II) for the different composites. These data are also consistent with the nano rod-like shape of these materials. The catalytic activity for the selective oxidation of olens is due to the effect of the redox ability of the modifying cations in the goethite-based nanocomposites and base sites/acid sites ratio. The stability of the modied goethites is sensible due to the action of the isobutyric acid generated as a by-product. The composites will be studied in other catalytic reactions considering that such materials that act at room temperature are easy to synthetized and does not imply high costs. To improve the catalytic abilities, these will be calcined, and the new oxide-based composites will also be tested from this point of view.

Conflicts of interest
There are no conicts to declare.