Chemistry Africa

, Volume 2, Issue 1, pp 77–87 | Cite as

Characterization and Catalytic Activity of Mn(salen) Supported on a Silica/Clay-Mineral Composite: Influence of the Complex/Support Interaction on the Catalytic Efficiency

  • Tesnime Abou Khalil
  • Souhir BoujdayEmail author
  • Juliette Blanchard
  • Latifa BergaouiEmail author
Original Article


We investigated different strategies to prepare Mn(salen) complexes supported on a silica/clay-mineral composite. The silica/clay-mineral surfaces were modified by grafting aminosilane groups, then Mn(salen) complexes were either grafted through a cross-linker, or deposited on the surface through interactions between amino group and manganese center. Two additional strategies, based on hydrogen bonds and electrostatic interactions, were also considered for comparison. The resulting materials were characterized using IR in diffuse reflectance mode, thermogravimetric analysis and elemental analysis in order to determine the manganese content and to establish the nature of the interaction between Mn(salen) complexes and silica/clay-mineral material surface for the different preparation protocols. These catalysts, were used for cyclohexene oxidation by tert-Butyl hydroperoxide with the objective of correlating the nature of the interaction between the metallic complex and the support to the catalytic activity. They were also compared to similar systems prepared using fumed silica as support, evidencing the input of this silica/clay-mineral composite in the catalytic performances. The obtained materials exhibited differences in both reactivity and selectivity that can be related to the nature of the interaction between the metallic center and the oxide support. The highest conversions were observed when the Mn complexes were immobilized with rigidness on the support. Interestingly, the resulting materials exhibited a remarkable stability, as evidenced by the absence of leaching of Mn(salen) in the reaction batch, and hence an excellent reusability.

Graphical Abstract


Mn(salen) Silica/clay-mineral material Functionalization Oxidation Cyclohexene 

1 Introduction

The epoxidation of unfunctionalized olefins in presence of Manganese(III) salen complexes acting as catalysts is attracting growing interest since 1990 [1, 2]. Mn(salen) complexes were extensively studied for this reaction in homogeneous [3, 4, 5] conditions. However, the high cost of chiral salen transition metal complexes (TMC) in addition to the difficulty to separate and recover these catalysts from the reaction mixture, led to a significant effort to develop heterogeneous Mn(salen) catalysts by immobilizing these complexes on support [6, 7].

For these systems, the choice of the support is crucial to achieve a successful immobilization. Various solid supports, especially inorganic or hybrid inorganic–organic porous materials like zeolites [8], silica [9, 10], carbon [11], clay minerals [12], and organosilica [13] as well as layered solids [14], were used. Inorganic supports such as clay minerals are environmentally more acceptable and more resistant to organic solvents than organic supports. In this context, the smectites are good candidates to serve as supports. These layered clay minerals have an important cation exchange capacity [15] due to their negatively charged layers [16] and their high swelling potentials [17]. However, their thermal stability is limited making the internal surface inaccessible and therefore offering very limited area for TMCs immobilization. To overcome this issue, the clay layers can be pillared using thermally stable pillars that prevent the interlayer spaces from collapsing (so-called PILCs materials) [18]. However, the microporous size of the interlayer space limits the diffusion of voluminous molecules to the active sites, and therefore, the resulting systems exhibit low catalytic activity. Consequently, engineering materials with a controlled porosity is needed to solve the difficulties of diffusion in both adsorption and catalysis processes. To this end, we recently engineered, through a sol–gel process, an amorphous mesoporous silica/clay mineral composite materials with controlled porosity [19]. In this work, these silica/clay-mineral composite materials are used as support to immobilize the Mn(salen) complex using different attachment methods with two objectives: first, improving the catalytic efficiency of the supported Mn(salen) complexes compared to conventional supports, and second, investigating the effect of the type of interaction between the support and the complexes on the catalytic properties.

The immobilization of chiral manganese (III) salen complexes can be achieved either by covalent bonds or by non-covalent interactions with the surface. Grafting and anchoring terms refer to the chemisorption of TMCs to the surface and thus exclude physisorbed species (through either hydrogen bonds or Van der Waals interactions). In the case of TMCs grafting, one or several groups of the support’s surface are chemically bound to the metal; the support is part of the metal inner coordination sphere and thus acts as a ligand. In the case of anchoring, the metal center of an organometallic complex is connected to the surface via one of its ligand that acts as linker (also-called tether or spacer). This attachment method has two advantages: first, it broadens the range of possible surface chemistries and second, it provides additional flexibility to the anchored complex [20]. In this paper we followed different strategies to deposit the Mn(salen) complexes on silica/clay-mineral composite surfaces. Mn(salen) and/or the aryldiamine Mn(salen) complexes were either physisorbed on silica/clay-mineral material or chemically attached to a composite modified by grafting aminopropyltriethoxysilane [21] (APTES) on its surface. In a last strategy, APTES-grafted composite was further modified by PDC (1,4-phenylenediisothiocyanate), acting as a cross-linker [22]. With these strategies, four catalytic materials were prepared as illustrated in Fig. 1 and carefully characterized in order to investigate the nature of the interactions between Mn(salen) complexes and silica/clay-mineral surface. These materials were then used in the cyclohexene oxidation reaction and their catalytic activities were measured and compared to those of Mn(salen) and aryldiamine Mn(salen) complexes tested in the same reaction as homogeneous catalysts. The catalytic performances of the systems synthesized in this work were compared to the results obtained previously using fumed silica as support [10]. The efficiencies are also discussed based on the preparation strategy and correlated to the nature of the link between Mn(salen) complexes and the silica/clay-mineral support.
Fig. 1

Modification procedure of Mn(salen) complex and the methodologies adopted to immobilize the manganese complex on silica/clay support, together with the name used for each catalyst

2 Materials and Methods

2.1 Materials

Tetraethoxysilane (TEOS, 99%, Aldrich), cetyltrimethylammonium bromide (CTAB, 99%, Sigma), hydrochloric acid (HCl, 37%, Scharlau), 3-aminopropyltriethoxysilane (APTES, 99%, Aldrich), (R,R)-(-)N,N-bis(3,5-di-tert-butylsalicylidene)-1,2-cyclohexane diaminomanganese (III) chloride ([MnC], Sigma-Aldrich), 1,4 phenylene diisothiocyanate (PDC, 98%, Sigma-Aldrich), tetrahydrofuran (THF, 99.9%, Sigma-Aldrich), p-phenylene diamine (Sigma) cyclohexene (99%, Sigma-Aldrich) and tert-butyl hydroperoxide (TBHP, 5.5 mol L−1 in decane, Sigma) were used without further purification.

2.2 Methods

2.2.1 Preparation of Silica/Clay-Mineral Composite

The synthesis and characterization of the silica/clay-mineral composite has been reported in a previous work [19]. Briefly, 17.75 ml of a TEOS/ethanol solution (1 mol L−1) were mixed with 1.5 g of CTMA-montmorillonite in presence of 1.44 ml of an aqueous HCl solution (0.8 mol L−1) until gelation. The use of a pre-swelled clay mineral provides a better delamination of the layers. The obtained gel was dried at 70 °C and then calcined at 600 °C under oxygen. During this calcination step, the organic component was decomposed and the clay mineral layers were totally dehydroxylated leading to a nearly amorphous solid. The surface area of this calcined silica/clay-mineral composite (hereafter denoted ├Si/clay) was about 381 m2 g−1 and this material was shown to contain both micropores and mesopores (Smeso = 184 m2 g−1; Smicro = 197 m2 g−1) [19].

Synthesis of modified Mn(III)salen [NH2MnC]: [MnC] was modified following the procedure described by Huang et al. [23, 24]: 1 mmol of Mn(III) salen was mixed with 10 mL of tetrahydrofurane, then 1 mmol of p-phenylene diamine was added. The mixture was kept under stirring with a reflux condenser for 15 h at 60 °C.

Preparation of 3-aminopropyl-modified silica/clay-mineral├Si/clay^^NH2: 5 g of ├Si/clay were first dispersed in 90 mL of dry toluene. After adding 6.8 mL of 3-(aminopropyl)-triethoxysilane (APTES), the suspension was stirred at 60 °C for 24 h and then filtered. The obtained solid was washed in a Soxhlet apparatus with methanol (in order to remove unreacted APTES) and finally dried at 80 °C for 4 h.

Preparation of modified silica/clay-mineral with the crosslinker├Si/clay^^NH-R-NCS: 1 g of APTES-modified composite (├Si/clay^^NH2) was added to a 20 mL of a 10.4 mmol L−1 solution of 1,4-phenylenediisothiocyanate in pyridine/DMF (10%/90%, v/v). The suspension was left under stirring during 2 h at room temperature and away from light [25]. The solid was washed with DMF then with ethanol and dried overnight at 75 °C.

Immobilization of Mn(III) salen complexes: Four catalysts were prepared following the procedures described in Fig. 1. For all samples, the solid support was mixed with the manganese complex solution during 7 h at room temperature and then filtered. The resulting material was washed by dichloromethane and dried overnight at 60 °C. To prepare ├OH + [NH2–MnC],├NH2 + [NH2–MnC] or ├NCS + [NH2–MnC], 0.5 g of ├Si/clay, ├Si/clay^^NH2 or ├Si/clay^^NH-R-NCS, respectively were mixed with 10 mL of the [NH2MnC] solution described above. For ├NH2 + [MnC] preparation, 1 g of ├Si/clay^^NH2 in 10 mL of dichloromethane was treated with 1 mmol of Mn(salen) complex.

2.2.2 Characterization

A Bomem MB 155 infrared spectrometer was used for IR in DRIFT mode measurements. After dissolution of the solid, the manganese was analysed by ICP using a Horiba Jobin–Yvon (Activa) apparatus. Carbon and nitrogen were analyzed by an EMIA-V2-Horiba 200 V analyzer. TG measurements were performed under air flow using a SDTQ600 apparatus. Samples were heated starting from 25 °C to 150 °C with a heating rate of 10 °C min−1. The temperature was then kept constant for 70 min at 150 °C and finally increased to 800 °C with a heating rate of 5 °C min−1.The temperature range between 25 and 150 °C corresponds to departure of water and the temperature range between 150 and 800 °C, corresponds to the decomposition of organic compounds. The organic matter loss was expressed on a dry weight basis.

2.2.3 Cyclohexene Oxidation

For the catalytic reactions, 5.0 mL of dichloromethane, 0.25 mL of cyclohexene, 0.1 mL of toluene (internal standard) and 0.05 g of catalyst were stirred at 40 °C in a 50 mL-round-bottom flask set with a reflux condenser and under N2. Then, the oxidant, 0.5 mL of t-BuOOH, was added to the stirred solution under nitrogen. After a reaction time of 24 h, products were collected and analysed using an Agilent GC gas chromatography equipped with an HP5-MS capillary column. A good separation of the analytes was obtained thanks to an appropriate temperature program and assignment of GC peaks to analytes was performed by comparison with retention time of commercial chemicals (see supporting information for temperature program of GC oven and retention time). The GC response for each analyte was calibrated by measuring known concentrations of the corresponding commercial chemical. Toluene was used as internal standard to correct from small variations in the injected volume (1 µL). A blank test (without catalyst) was performed under the same conditions to establish the absence of oxidation of cyclohexene under these conditions and a very low conversion (below 1%) was indeed observed.

To validate the heterogeneity of the catalytic process and the reusability or the catalyst, the catalyst was separated from the reaction medium after a reaction time of 24 h, and then the reaction was continued for additional 24 h. The recovered catalyst was dried at 60 °C and re-used in the cyclohexene oxidation reaction.

3 Results and Discussion

Mn(salen) complexes ([MnC]) were deposited on the silica/clay-mineral support (├Si/clay) following four strategies leading to different types of interaction between the manganese complex and the support as depicted in Fig. 1. First, an aryldiamine modified [MnC] was prepared by adding a phenylenediamine [22] ligand to the metal center ([NH2–MnC]). The first catalytic system, ├OH + [NH2-MnC], was obtained by physisorbing this complex on the silica/clay-mineral material. Two catalytic materials were obtained by reacting [MnC] or the aryldiamine modified Mn(salen), [NH2–MnC] on the APTES modified composite,├Si/clay^^NH2 support: ├NH2 + [NH2–MnC] and ├NH2 + [MnC], respectively. Finally, the PDC cross linker was grafted to the APTES modified composite and then used to covalently bind [NH2–MnC] and form the fourth catalytic system, ├NCS + [NH2–MnC]. For the two systems ├OH + [NH2–MnC] and├NH2 + [NH2–MnC], the interaction with the support should not involve covalent or coordination bond formation; only electrostatic interactions and/or H-bonds are expected. For the two other systems ├NH2 + [MnC] and ├NCS + [NH2–MnC], we expect covalent bond formation, either through a direct reaction between the terminal amine group of├Si/clay^^NH2 and the metal centre for ├NH2 + [MnC] or through reaction between the amine ligand of [NH2–MnC] complex and the PDC group of├Si/clay^^NH-R-NCS for ├NCS + [NH2–MnC]. Before assessing the catalytic activity of the so-prepared materials, IR, TG and elemental analysis were used to validate the assumed mechanism of interaction for each sample. In what follows we present and discuss these data.

3.1 Silica/Clay-Mineral Support Functionalization and Characterization

Silica/clay-mineral was used as support for Mn complexes either pristine, or modified by APTES and eventually by coupling the cross linking agent, PDC, to the amine groups of ├Si/clay^^NH2. IR spectroscopy in DRIFT mode was used to follow these successive steps. The spectra obtained for ├Si/clay, ├Si/clay^^NH2 and ├Si/clay^^NH-R-NCS are shown in Fig. 2a–c, respectively. All spectra show some common features arising from the silica/clay-mineral network: a band at 1288 cm−1, assigned to hydroxyl groups bending, a second band at 798 cm−1 attributed to siloxane Si–O–Si stretching [26], the stretching of the OH groups from silanols involved in hydrogen bonds leads to a very large band around 3400 cm−1 and includes water molecule vibrations, whereas the narrower band observed at 3630 cm−1 is assigned to isolated OH groups [27, 28], and, finally, the small band around 1625 cm−1 is attributed to adsorbed water molecules.
Fig. 2

IR spectra in DRIFT mode of a ├Si/clay, b ├Si/clay^^NH2 and c ├Si/clay^^NH-R-NCS, d ├OH + [NH2-MnC], e ├NH2 + [NH2-MnC], f ├NH2 + [MnC] and g ├NCS + [NH2-MnC]. The multiplication coefficient, that has been applied to the spectra in the 4000–2500 cm−1 range to facilitate their comparison, is indicated on the left side

After APTES grafting (sample ├Si/clay^^NH2), the IR spectrum (Fig. 2b) shows additional bands, with the typical bands corresponding to C–H vibrations of aliphatic groups appearing at 2923 and 2863 cm−1 [29, 30] and a broad band 3250 cm−1 that can be assigned to N–H vibrations [25, 31]. These bands confirm the presence of APTES on the surface.

After reaction with the PDC (├Si/clay^^NH-R-NCS sample), the IR spectrum is clearly different from that observed for ├Si/clay^^NH2 (Fig. 2c). First, an intense band that can be associated to C=N stretching of N=C=S group appears at 1658 cm−1. The vibration band of –N–C=S [32, 33] is also observed at 1389 cm−1. In addition, a shoulder appears at 3052 cm−1 that can be ascribed to the aromatic C–H vibration [34]. Finally, two additional, weak but sharp, bands are observed at 1411 cm−1 and 1434 cm−1 and are assigned to C–N stretching frequency [35] and to the aromatic C=C [36], respectively. All these features confirm the presence of the PDC molecule and the band associated with the –N–C=S group indicates that the PDC molecule has reacted with the NH2 terminal group of APTES, whereas the band associated with C=N=S group indicates the availability of this group for further reactions. However, the persistence of ʋ(NH) vibrations, whose intensity increases for ├Si/clay^^NH-R-NCS compared to ├Si/clay^^NH2 suggests an incomplete reaction of the cross-linker on the amine groups.

Thermogravimetric analysis was also used to investigate the modification of the silica/clay-mineral support and estimate the coverage for the modified support. The mass losses calculated from TG data are gathered in Table 1.
Table 1

TG mass losses and chemical analysis of the functionalized solids


% Mass loss

Elements (mmol g−1)

Ratio mol/mol

20–150 °C

150–800 °C*




















Mn catalysts

 ├OH + [NH2–MnC]







40 (42)

3.8 (4)

 ├NH2 + [NH2–MnC]







53 (45)

12.8 (5)

 ├NH2 + [MnC]







60 (39)

9.8 (3)

 ├NCS + [NH2–MnC]







73 (53)

15.9 (7)

*TG mass loss expressed on a dry weight basis. Value between brackets refers to theoretical value

The results displayed in this table are expressed for two temperature ranges 20–150 °C corresponding to the departure of physisorbed water and 150–800 °C associated with organic matter decomposition. In the first region, there are no significant differences in mass losses between ├Si/clay^^NH2 and ├Si/clay^^NH-R-NCS suggesting similar amounts of physisorbed water molecules and therefore, similar hydrophilic character. For the higher temperature range 150–800 °C, the mass loss was clearly higher for ├Si/clay^^NH-R-NCS in agreement with the addition of PDC cross-linker to the surface. The amount of APTES present on ├Si/clay^^NH2 surface was estimated from this data to 1.5 mmol g−1. This value may seem high compared to the results of Pereira et al. [37], i.e. 0.9 mmol g−1 for APTES-laponite support and 1.1 mmol g−1 for APTES-K10-montmorillonite, but is very close to the previous results found with silica [10]. Note that the value 1.5 mmol g−1 was obtained assuming that the organic loss corresponds exclusively to the departure of –C3H6–NH2 without considering the presence of the few residual ethoxy groups observed on the IR spectra. The concentration would be slightly lower (but still higher than 1 mmol g−1) if, in addition to –C3H6–NH2, we consider the departure of one ethoxy group per grafted APTES molecule, which is clearly an overestimation, considering the low intensity of the IR bands at 1124 and 1203 cm−1 in Fig. 2b.

TG results show that the weight loss due to PDC decomposition is about 3.9%, i.e. 0.2 mmol of PDC per gram of solid, a value that is relatively low compared to the APTES coverage measured above. Thus, only 13 to 20% of the APTES extremities have reacted with the crosslinker and the surface is terminated by more amine functions than thiocyanates.

To further investigate silica/clay-mineral support modification, elemental analyses were conducted. The results are summarized in Table 1. For ├Si/clay^^NH2, the C and N elemental analyses show that the ratio C/N is slightly higher than 3 (precisely 3.2). This value is consistent with the structure of a grafted APTES molecules; the small deviation of 0.2 may result from the persistence of few ethoxy groups on the surface. On ├Si/clay^^NH-R-NCS, the addition of PDC, which contains two terminal ends N=C=S groups, leads to an increase of both N and C content. The amount of PDC linked to the surface can be estimated from the increase in N concentration compared to ├Si/clay^^NH2: the nitrogen coverage from PDC molecule (which contains 2 N) is 0.42 mmol g−1, thus PDC coverage is 0.21 mmol g−1. Therefore, consistently with the data obtained from thermal analysis, only 20% of APTES molecules were further modified by PDC and the support referred to as ├Si/clay^^NH-R-NCS is terminated by a mixture between –N=C=S and –NH2 with a large dominance of amine groups.

3.2 Supported Mn(salen) Catalysts Characterization

IR spectra in DRIFT mode of the supported Mn catalysts are shown in Fig. 2. ├OH + [NH2–MnC] (Fig. 2d), ├NH2 + [NH2–MnC] (Fig. 2e), ├NH2 + [MnC] (Fig. 2f) and ├NCS + [NH2–MnC] (Fig. 2g) samples present the same large and poorly resolved spectra. It is worth noting, for all spectra, the low intensity of the band at 1288 cm−1 assigned to hydroxyl group. The comparison of ├NH2 + [NH2–MnC] and ├NH2 + [MnC] spectra (Fig. 2e–f, respectively) to ├Si/clay^^NH2 one, (Fig. 2b) shows the persistence of a broad band associated with υ(NH) vibration at ca. 3350 cm−1 of the grafted APTES molecules indicating that not all amine groups were consumed upon adding Mn complexes. Besides, the spectrum shows new bands at 1610 and 1530 cm−1, assigned to C=N and C=C groups of the salen ligand, respectively, which confirms their presence on the surface.

The comparison of ├NCS + [NH2–MnC] spectrum (Fig. 2g) to ├Si/clay^^NH-R-NCS one (Fig. 2c) shows that the intensity of the band at 1389 cm−1 related to –N–C=S is still present. The band assigned above to the C=N vibration is also still present, but shifts from 1658 to 1648 cm−1. The presence of this band suggests the incomplete reaction between the free end of the crosslinker and the amine group belonging to the metal complex. The ├NCS + [NH2–MnC] spectrum (Fig. 2g) illustrates also the presence of a shoulder at 3062 cm−1 assigned to aromatic C–H vibration. Finally, the bands in the range 3300–3500 cm−1 are ascribed to N–H vibrations of Mn(salen) in addition to the C=N and C=C bands.

Thermogravimetric analysis was used to evaluate the amount of retained Mn complexes. Table 1 shows the organic material contents between 150 and 600 °C for the four prepared samples. After immobilization of the Mn complex the mass loss between 150 and 800 °C did not considerably increase. These amounts were significantly lower than those obtained by similar strategies using a non-porous SiO2 support [10]. These lower loadings probably have two origins: first, in the composite material, the silanols are mostly located in the SiO2-like phase of the composite (the clay-mineral layers are expected to bear silanols only at their edges); second, the composite contains both micropores and mesopores and the bulkiness of the manganese complexes will likely impedes/limits its diffusion into the smaller pores of the silica/clay-mineral composite. The amounts of retained Mn determined from the TG measurements led to the following values: 0.10 mmol g−1 for ├OH + [NH2–MnC], 0.03 mmol g−1 for ├NH2 + [NH2–MnC], 0.04 mmol g−1 for ├NH2 + [MnC] and 0.01 mmol g−1 for ├NCS + [NH2–MnC].

For the silylated supports, if the metallic complex was immobilized via the grafted APTES molecule, only 8% for ├NH2 + [NH2–MnC], 10% for ├NH2 + [MnC] and 3% for ├NCS + [NH2–MnC] of APTES would be involved in this immobilization.

Chemical analysis can provide more precise information about the amount of retained manganese complexes. Results of Mn, C and N analyses are given in Table 1. For all supports, a manganese uptake ranging between 0.08 and 0.13 mmol g−1 is observed, indicating that, whatever the synthesis protocol Mn(salen) complexes interact strongly enough with the support to withstand the washing step. Moreover, the Mn(salen) loadings are significantly lower than the APTES loading (about 1.5 mmol g−1) and the PDC loading (about 0.21 mmol g−1). Hence, based on chemical analysis, the Mn(salen) loadings are not inconsistent with the theoretical immobilization schema (Fig. 1). The theoretical N/Mn and C/Mn molar ratios for the modified Mn complex (C42H59ClMnN4O2) are 4 and 42, respectively. The elemental analysis of ├OH + [NH2–MnC] leads to values equal to 3.8 and 40 for N/Mn and C/Mn molar ratio, respectively. The good agreement between experimental and theoretical values shows the excellent stability of the modified Mn complex during its interaction with the solid surface. Besides, the Mn amount of 0.13 mmol g−1 obtained for ├OH + [NH2–MnC] sample matches with the amount [NH2–MnC] determined by thermogravimetric analysis. For├NH2 + [NH2–MnC], N/Mn and C/Mn molar ratio are 13 and 53, respectively instead of 5 and 45 if all the grafted APTES molecules have reacted with the modified metallic complex. This result confirms that [NH2–MnC] amount is lower than APTES coverage. All the same, for ├NH2 + [MnC], the measured N/Mn and C/Mn ratio are equal to 10 and 60, respectively, instead of 3 and 39 if the reaction with surface aminogroups was total. In the case of the sample ├NCS + [NH2–MnC], the theoretical N/Mn and C/Mn ratio are 7 and 53, respectively. Experimental ratios are very different from the expected ones and are not in agreement with TG results. Moreover, DRIFT results did not confirm the existence of a covalent bond between the free side of PDC and manganese complex. Thus, at this stage, the strategy applied to synthesis the catalyst ├NCS + [NH2–MnC] is neither confirmed nor refuted by the characterization techniques. Note that, for these systems, ├NCS + [NH2–MnC] and for ├NH2 + [NH2–MnC], the amounts of Mn were similar. We can therefore not exclude, based on characterization solely that the mode of interaction of the Mn complex with the surface is the same for these two catalysts.

3.3 Catalytic Test

Catalytic activities and selectivities of the four catalytic materials prepared above were assessed for the oxidation of cyclohexene using tert-butyl hydroperoxide (TBHP) as the oxidant in the liquid phase. The products of this catalytic reaction are cyclohexenol (Enol), cyclohexenone (Enone), cyclohexanone (One), cyclohexene oxide (Epox) and cyclohexane diols (Diol). To obtain these products two pathways must be considered as illustrated in Fig. 3: The oxidation of the C=C bond of cyclohexene with peroxides yields the cyclohexene oxide which upon further reaction with water, produces cyclohexane-1,2-diol. The oxidation of the allylic C–H bond results in alcohol (2-cyclohexene-1-ol in equilibrium with cyclohexanone) and ketone product (2-cyclohexene-1-one). The resulting products, 2-cyclohexene-1-ol and 2- cyclohexene-1-one, are used in the manufacture of high-value pharmaceuticals [38], and cyclohexene oxide can react in cationic polymerization to yield poly(cyclohexene oxide) which is useful in many applications (coatings, ink and adhesives) [39].
Fig. 3

Schematic representation of the two main pathways to the oxidation of cyclohexene

The catalytic results obtained for the four catalytic materials are summarized in Table 2.
Table 2

Cyclohexene oxidation conversion and selectivities of different catalysts after 24 h of reaction time


Conv. (%)


Conv. (mol/mol)a

Selectivity (%)

(mol g−1)






├OH + [NH2–MnC]









├NH2 + [NH2–MnC]









├NH2 + [MnC]









├NCS + [NH2–MnC]









[MnC]b, [10]









[NH2–MnC]b, [10]









SiO2 + [NH2–MnC] [10]









aNumber of moles of reacted cyclohexene per mole of manganese, determined from chemical analysis, in the batch after 24 h of reaction time

bHomogeneous catalysis

For comparison, the catalytic performances of the Mn(salen) and the aryldiamine Mn(salen) complexes, measured in the same conditions but in homogeneous mode, are also gathered in Table 2. For homogeneous catalysts, the conversions, when expressed as a function of manganese amount in the reactive media are similar for the two catalysts. Therefore, we can conclude that the addition of the aryldiamine ligand to the Mn(salen) does not affect the accessibility nor the efficiency of the active sites in homogenous catalysis.

For heterogeneous catalysts, the two functionalized composites prepared using the aryldiamine modified Mn(salen) [NH2–MnC] and based on a weak interaction with the support (├OH + [NH2–MnC] and ├NH2 + [NH2–MnC]) gave conversions equal to 37.6 and 23.6%, respectively. Here also, when these conversions are expressed as function of manganese amount, their values are similar but noticeably higher (more than five times) than the conversions obtained using the corresponding homogeneous catalyst.

The comparison of the conversion (expressed as a function of the manganese amount) obtained for ├OH + [NH2–MnC] to the catalyst prepared following the same strategy but using fumed silica as support [10] (noted SiO2 + [NH2–MnC] in Table 2) indicates that ├OH + [NH2–MnC] is slightly more efficient (~ 34% higher) showing a positive input of the silica/clay-mineral composite. The two supports have slightly the same surface area (390 and 381 m2 g−1 for the fumed silica and the silica/clay-mineral, respectively). Therefore, this enhancement may result from the porosity of the silica/clay-mineral composite compared to the non-porous fumed silica. When comparing the catalytic performances of the four systems prepared herein, the data from Table 2 show that the conversion (relative to the number of mol of manganese) varies by a factor of almost 2.5. Hence, the synthesis protocol has a clear incidence on the activity of the catalyst. This does not seem to be related to the MnC loading as catalysts with similar loading (e.g. ├NCS + [NH2–MnC] and ├NH2 + [NH2–MnC] show very distinct conversions (per manganese).

In terms of conversion per manganese, the catalysts can be separated in two to three groups: in the first group, the two catalysts with the lowest conversions (144.4 and 146.3 mol of converted cyclohexene per mol of manganese in 24 h): ├OH + [NH2–MnC] and ├NH2 + [NH2–MnC]; in the second and third group, the two catalysts with the highest conversion (290.7 and 357.4 mol/mol): ├NCS + [NH2–MnC] and ├NH2 + [MnC]. Interestingly, this classification into two with distinct conversion is linked to the strength of the interaction between the metal and the support: the first group corresponds to the catalysts for which a weak interaction on the MnC with the support is expected, whereas, in the second group, a stronger interaction is expected: for ├NH2 + [MnC], one expect that the NH2 group enter in the coordination sphere of the manganese, whereas, for ├NCS + [NH2–MnC] the manganese complex is expected to be tethered to surface by the reaction of the terminal NCS group with the amine group of [NH2–MnC] (see Fig. 1). This relationship between immobilization and activity can be extended to the homogeneous catalysts: indeed the weaker activity of these catalysts compared to their heterogeneized counter parts could be assigned to their higher mobility.

[MnC] ≈ [NH2–MnC] ≪ ├OH + [NH2–MnC] ≈ ├NH2 + [NH2–MnC] ≪ ├NCS + [NH2–MnC] < ├NH2 + [MnC]. Moreover, the slightly higher activity of ├NH2 + [MnC] compared to ├NCS + [NH2–MnC], could reflect the stronger immobilization (shorter tethering) of├NH2 + [MnC].

The differences in selectivities may result from the considerable differences in the catalytic conversions, and therefore their comparison is not straightforward. But a general observation can me made for the different catalytic pathways: for the catalysts prepared starting from [NH2–MnC] modified complex the allylic oxidation was predominant and the main product was the Enone (~ 47%), consistently with the results for the [NH2–MnC] complex in homogeneous catalysis. Therefore, we can assume that the catalytic pathway is governed by the nature of the transition metal complex rather than the interaction with the support, yet, thus must be validated for identical catalytic conversions.

Finally, we assessed the heterogeneity of the catalytic reaction by recovering the catalyst and continuing the reaction to ensure that there were no dissolved Mn complexes. Therefore, after 24 h of reaction, the catalysts were separated from the liquid phase and the reaction was allowed to continue in the homogeneous mixture. For all the four systems, no additional conversion of cyclohexene was observed. This clearly indicates that the catalytically active centers are not subject to leaching phenomena. Moreover, the study of recyclability of composite catalysts during the cyclohexene oxidation after two cycles was performed (Fig. 4).
Fig. 4

Catalytic conversion of cyclohexene oxidation for two successive runs

The results obtained by comparing two successive runs show no decrease in the conversions. These features confirm the stability of the active sites and the absence of leaching of the Mn complexes under reaction conditions, independently of the synthesis strategy. In addition, when fumed silica was used as a support, Mn-catalysts showed lower chemical stability since a small decrease in the activity (2–5%) during the second run was observed [10]. With this composite silica/clay-mineral material the catalytic activity totally preserved between the first and the second cycle. The porosity of this silica/clay-mineral material, compared to fumed silica porosity, is a possible explanation to this enhancement as it may protect the active sites and further stabilize the catalytic system.

4 Conclusion

Mn(salen) complexes supported on a silica/clay-mineral composite material were prepared following different strategies aiming either at an anchoring of the complex to the surface or at weaker interactions (electrostatic or Van der Waals type interactions). Our approaches included the modification of the metallic complex (by adding an aryldiamine ligand) and/or the grafting of functional molecules on the composite surfaces. To avoid active species leaching, the loosely bound Mn(salen) complexes were removed by washing. The combination of several characterization techniques allowed us to validate three strategies among the four explored herein. However, for the strategy involving a cross linker on the modified Mn(salen) complexes, the establishment of a covalent bond was not demonstrated by our spectroscopic data but remains possible. The use of the silica/clay-mineral composite as support led to a considerable improvement of the catalytic performances of the systems synthesized in this work compared to the results obtained using fumed silica as support. In addition, the catalytic results showed that modifying the Mn(salen) by adding the aryldiamine ligand did not affect its activity in the cyclohexene oxidation but had an influence on its selectivity. We also observed that the catalytic properties were enhanced when the catalytic sites were immobilized with rigidness on the support; the highest performances were recorded for the strategy involving a silylation of the support followed by the Mn complex immobilization through a grafting mechanism. Regardless of the strategy employed for the catalysts synthesis, the interaction between the active site and the composite surface was irreversible allowing the reusability of the catalysts without any leaching. This stability was slightly improved compared to fumed silica confirming the considerable input arising from the use of these silica/clay-mineral composites porous supports for these systems.


Compliance with ethical standards

Conflict of interest

On behalf of all authors, the corresponding author states that there is no conflict of interest.

Supplementary material

42250_2018_23_MOESM1_ESM.docx (22 kb)
Supplementary material 1 (DOCX 21 kb)


  1. 1.
    Zhang W, Loebach JL, Wilson SR, Jacobsen EN (1990) J Am Chem Soc 112:2801CrossRefGoogle Scholar
  2. 2.
    Irie R, Noda K, Ito Y, Matsumoto N, Katsuki T (1990) Tetrahedron Lett 31:7345CrossRefGoogle Scholar
  3. 3.
    Jacobsen EN, Zhang W, Güler ML (1991) J Am Chem Soc 113:6703CrossRefGoogle Scholar
  4. 4.
    Maity NC, Abdi SHR, Kureshy RI, Khan NH, Suresh E, Dangi GP, Bajaj HC (2011) J Catal 277:123CrossRefGoogle Scholar
  5. 5.
    Tan R, Yin D, Yu N, Jin Y, Zhao H, Yin D (2008) J Catal 255:287CrossRefGoogle Scholar
  6. 6.
    Garcia AM, Moreno V, Delgado SX, Ramírez AE, Vargas LA, Vicente MÁ, Gil A, Galeano LA (2016) J Mol Catal A: Chem 416:10–19CrossRefGoogle Scholar
  7. 7.
    Fraile JM, García JI, Herrerías CI, Mayoral JA, Pires E (2009) Chem Soc Rev 38:695CrossRefGoogle Scholar
  8. 8.
    Salavati-Niasari M, Ganjali MR, Norouzi P (2007) J Porous Mater 14:423CrossRefGoogle Scholar
  9. 9.
    Hatefi M, Moghadam M, Sheikhshoaei I, Mirkhani V, Tangestaninejad S, Mohammad Poor-Baltork I, Kargar H (2009) Appl Catal 370:66CrossRefGoogle Scholar
  10. 10.
    Abou Khalil T, Ben Chaabene S, Boujday S, Bergaoui L (2015) IOSR-JAC 8:36Google Scholar
  11. 11.
    Salavati-Niasari M, Badiei A, Saberyan K (2011) Chem Eng J 173:651CrossRefGoogle Scholar
  12. 12.
    Kuzniarska-Biernacka I, Pereira C, Carvalho AP, Pires J, Freire C (2011) Appl Clay Sci 53:195CrossRefGoogle Scholar
  13. 13.
    Hu JL, Wu QY, Li KX, Li W, Ma FY, Zhang SQ, Su F, Guo YH, Wang YH (2010) Catal Commun 12:238CrossRefGoogle Scholar
  14. 14.
    Parida KM, Singha S, Sahoo PC (2010) J Mol Catal A: Chem 325:40CrossRefGoogle Scholar
  15. 15.
    Veld B (1992) Introduction to clay minerals. Springer, NetherlandsCrossRefGoogle Scholar
  16. 16.
    Newman ACD (1987) Chemistry of clays and clay minerals. Wiley, New YorkGoogle Scholar
  17. 17.
    Bergaya F, Theng BKG, Lagaly G (2006) Handbook of clay science, 1st edn. Elsevier, AmsterdamGoogle Scholar
  18. 18.
    Gil A, Korili SA, Trujillano R, Vicente MA (2010) Pillared clays and related catalysts. Springer, New YorkCrossRefGoogle Scholar
  19. 19.
    Abou Khalil T, Ben Chaabene S, Boujday S, Blanchard J, Bergaoui L (2015) J Sol-Gel Sci Technol 75:436CrossRefGoogle Scholar
  20. 20.
    Valkenberg MH, Hölderich WF (2002) Catal Rev 44:321CrossRefGoogle Scholar
  21. 21.
    Moller K, Bein T (1998) Chem Mater 10:2950CrossRefGoogle Scholar
  22. 22.
    Huang J, Fu X, Wang G, Ge Y, Miao Q (2012) J Mol Catal A: Chem 357:162CrossRefGoogle Scholar
  23. 23.
    Huang J, Fu X, Wang G, Ge Y, Miao Q (2012) Catal Sci Technol 2:1040CrossRefGoogle Scholar
  24. 24.
    Huang J, Fu X, Miao Q (2011) Appl Catal A 407:163CrossRefGoogle Scholar
  25. 25.
    Salmain M, Ghasemi M, Boujday S, Pradier C-M (2012) Biosens Bioelectron 173:148Google Scholar
  26. 26.
    Bois L, Bonhommé A, Ribes A, Pais B, Raffin G, Tessier F (2003) Colloids Surf A 221:221CrossRefGoogle Scholar
  27. 27.
    Meera KMS, Sankar RM, Murali A, Jaisankar SN, Mandal AB (2012) Colloids Surf B 90:204CrossRefGoogle Scholar
  28. 28.
    Qian Z, Hu G, Zhang S, Yang M (2008) Phys B 403:3231CrossRefGoogle Scholar
  29. 29.
    Silva AR, Botelho J (2014) J Mol Catal A: Chem 381:171CrossRefGoogle Scholar
  30. 30.
    Al-Oweini R, El-Rassy H (2009) J Mol Struct 919:140CrossRefGoogle Scholar
  31. 31.
    Silva AR, Wilson K, Clark JH, Freire C (2006) Microporous Mesoporous Mater 91:128CrossRefGoogle Scholar
  32. 32.
    Rao CNR, Venkataraghavan R (1962) Spectrochim Acta 18:54Google Scholar
  33. 33.
    Bedford EE, Boujday S, Humblot V, Gu FX, Pradier C-M (2014) Colloids Surf B 116:489CrossRefGoogle Scholar
  34. 34.
    Boujday S, Nasri S, Salmain M, Pradier C-M (2010) Biosens Bioelectron 26:1750CrossRefGoogle Scholar
  35. 35.
    Luo R, Tan R, Peng Z, Zheng W, Kong Y, Yin D (2012) J Catal 287:170CrossRefGoogle Scholar
  36. 36.
    Kim A, Filler MA, Kim S, Bent SF (2005) J Am Chem Soc 127:6123CrossRefGoogle Scholar
  37. 37.
    Pereira C, Patrício S, Silva AR, Magalhães AL, Paula Carvalho A, Pires J, Freire C (2007) J Colloid Interface Sci 316:570CrossRefGoogle Scholar
  38. 38.
    Cai ZY, Zhu MQ, Chen J, Shen YY, Zhao J, Tang Y, Chen XZ (2010) Catal Commun 12:197CrossRefGoogle Scholar
  39. 39.
    Dietliker K (1991) Chemistry and technology of UV and EB formulation for coatings, inks and paints, vol III. SITA Technology, LondonGoogle Scholar

Copyright information

© The Tunisian Chemical Society and Springer Nature Switzerland AG 2018

Authors and Affiliations

  1. 1.Laboratoire de Chimie des Matériaux et Catalyse, Faculté des Sciences de TunisUniversité Tunis El ManarTunisTunisia
  2. 2.Laboratoire de Réactivité de SurfaceCNRS, LRS UMR 7197, Sorbonne UniversitéParisFrance
  3. 3.Department of Biological and Chemical EngineeringCentre Urbain Nord, INSAT, Carthage UniversityTunisTunisia

Personalised recommendations