N,N-dimethylethylenediamine (12.6 mL, 115.3 mmol) was added to 3-isocyanatopropyl-(trimethoxy)silane (23.7 g, 115.3 mmol) and triethylamine (24 mL, 172 mmol) dissolved in 200 mL dry dichloromethane. The resulted mixture was stirred at room temperature for 24 h, and then it was concentrated by evaporation. 32.6 g of bright yellow liquid was obtained (yield 96.6%). ¹H NMR (300 MHz, CDCl₃): δ 0.66 (t, 7.8 Hz, 2H), 1.61 (q, 8.7 Hz, 2H), 2.25 (s, 6H), 2.426 (t, 5.7 Hz, 2H), 3.16 (q, 6 Hz, 2H), 3.25 (q, 6.9 Hz, 2H), 3.57 (s, 9H), 5.0 (s, 1H), 5.1 (s, 1H). ¹³C NMR (300 MHz, CDCl₃): δ 6.8, 23.8, 38.4, 43.3, 45.6, 50.9, 59.4, 159.1. IR (cm⁻¹): 3312, 2939, 2840, 2772, 1637, 1554, 1462, 1081. Elemental Analysis: calculated for C₁₁H₂₇N₃O₄Si, C: 45.02, H: 9.27, N: 14.32; found C: 44.47, H: 9.41, N: 14.01.

Thirty-five milliliters (192 mmol) of (3-chloropropyl)trimethoxysilane and 25.2 mL (192 mmol) of 1-butylimidazole were heated under N₂ at 120 °C. The progress of the reaction was monitored by ¹H NMR, and the reaction was completed after 8 h. The mixture was cooled to room temperature and 60.4 g of honey-like, orange viscous liquid was obtained (97% yield). ¹H NMR (300 MHz, CDCl₃):δ 0.616–0.672 (m, 2H), 0.967 (t, J = 7.5 Hz, 3H), 1.352 (dt, J = 8.1 Hz, 2H), 1.872–2.075 (m, 4H), 3.571 (s, 9H), 4.334 (dt, J = 4.2 Hz, 4H), 7.45 (t, J = 1.2 Hz, 1H), 7.568 (t, J = 1.8 Hz, 1H), 10.796 (s, 1H). ¹³C NMR (300 MHz, CDCl₃): δ 5.94, 13.48, 19.51, 24.21, 32.21, 49.79, 50.75, 51.7, 121.91, 122.07, 137.84. IR (cm⁻¹): 3128, 3044, 2943, 2843, 1560. Elemental Analysis: calculated for C₁₃H₂₇ClN₂O₃Si, C: 48.35, H: 8.43, N: 8.68; found C: 48.78, H: 8.78, N: 9.05.

Eleven point six grams of FeCl₃·6H₂O and 4.3 g of FeCl₂·4H₂O were mixed in 400 mL of degassed water. The mixture was heated at 85 °C, and then 15 mL of concentrated ammonia (28%) were added quickly. The heating was continued for another 30 min. After cooling to room temperature, the solution was decanted and the black precipitate was washed five times with 200 mL water. The black magnetite nanoparticles were suspended in 500 mL of ethanol (95%) and sonicated for 60 min. The resulted suspension was mechanically stirred and a solution of 100 mL ethanol (95%) containing 30 mmol of 1-butyl-3-(3-trimethoxysilylpropyl)-1H-imidazol-3-ium chloride and 2 mL of concentrated ammonia (28%) were added. Stirring under N₂ was continued for 36 h. The modified magnetite nanoparticles were magnetically separated and washed three times with 100 mL of ethanol (95%) and then dissolved in 400 mL of methanol and stirred mechanically for 30 min. One hundred milliliters of ether were added and the modified nanoparticles were magnetically separated, washed with 100 mL of ether and dried under a vacuum of 0.2 mmHg for 24 h. Typically, 6–6.5 g of brownish-black powder could be obtained.

The magnetite nanoparticles were prepared following the procedure described in Section 2.3 and the same amounts of reagents were used. The prepared magnetite was dispersed in 500 mL ethanol (95%) and sonicated for 60 min. After sonication, the suspension was mechanically stirred and 46.2 mmol of 1-(2-(dimethylamino)ethyl)-3-(3-(trimethoxysilyl)propyl)urea or N-[3-(trimethoxysilyl)propyl]--ethylenediamine dissolved in 100 mL ethanol (95%) were added. The mixture was stirred under N₂ for 36 h. The magnetite nanoparticles were separated from the mixture by application of an external magnetic field, and they were washed twice with ethanol (95%) followed by washing three times with methanol. The washed magnetite nanoparticles were dispersed again in 100 mL methanol and the resulted suspension was sonicated for 30 min. The concentration of the magnetite nanoparticles functionalized with amine groups was 50–54 mg/mL. Before any use, the suspension was sonicated for 10 min.

Two hundred milligrams of magnetic nanoparticles modified with 1-butyl-3-(3-trimethoxysilylpropyl)-1H-imidazol-3-ium chloride were dispersed in 5 mL methanol. This suspension was added to tetraethoxysilane (6 mL, 27.1 mmol) and distilled water (3.2 mL, 177.8 mmol). Then, 0.05 mL of HCl (0.1 N) were added and the mixture was stirred at room temperature for 2 h. After that, 4.6 mmol of 1-(2-(dimethylamino)ethyl)-3-(3-(trimethoxysilyl)propyl)urea or N-[3-(trimethoxysilyl)propyl]-ethylenediamine dissolved in 2 mL methanol were added to the mixture. A gellation of the mixture occurred after stirring for 4 h and the wet gel was aged at room temperature for 16 h. The resulting gel was dried under a vacuum of 0.2 mmHg at 60 °C for 24 h. Typically, 2.4–2.7 g of black xerogel were obtained as a final product.

Magnetite nanoparticles containing 0.19 mmol of the desired amine group were separated magnetically from the methanol and re-dispersed directly in 5 mL of a proper solvent. Then, 1.64 mmol of the desired aldehyde and nitroethane (0.17 mL, 2.39 mmol) were added. The resulting mixture was heated at 80 °C for 16 h. After cooling the mixture to room temperature, the catalyst was separated by applying an external magnetic field, and then the solution was decanted and dried over MgSO₄. The solvent was evaporated and the resulting products were analyzed by spectral methods. When Dimethyl sulfoxide (DMSO) was used as a solvent for the catalytic reaction, it was mixed with 10 mL of water after separation of the catalyst and the products were extracted three times with 20 mL of ether. The combined solutions of ether were washed three times with 15 mL water and then dried over MgSO₄. The ether was evaporated and the products were analyzed by spectral methods.

One hundred milligrams of silica xerogels containing 0.19 mmol of the desired amine groups were dispersed in 5 mL of a proper solvent. 1.64 mmol of the desired aldehyde and mitroethane (0.17 mL, 2.39 mmol) were added to the mixture. The resulting mixture was heated at 80 °C for 16 h and then cooled to room temperature. After magnetic separation of the catalyst, the solution was decanted and dried over MgSO₄. The solution was concentrated and the resulting products were analyzed by spectral methods. When DMSO was used as a solvent, the products were isolated from the solvent in the same manner as it is described in Section 2.5.

As shown in Scheme 1, two base catalysts functionalized with a trimethoxysilane group were anchored to the magnetite nanoparticles by a condensation reaction under basic conditions between the trimethoxysilane group and the hydroxyl groups existing on the surface of the magnetite nanoparticles. The first base catalyst, 1-(2-(dimethylamino)ethyl)-3-(3-(trimethoxysilyl)propyl)urea, contains tertiary amine groups and was prepared by the reaction of 3-isocyanatopropyl(trimethoxy)silane and N,N-dimethylethylenediamine. The second base catalyst, N-[3-(trimethoxysilyl)propyl]ethylenediamine, contains primary and secondary amine groups.

After supporting the base catalysts on the magnetite nanoparticles the resulting materials MNP-NMe₂ and MNP-NH₂ were characterized by Transmission Electron Microscopy (TEM), X-ray Diffraction (XRD), thermogravimetric analysis-differential thermal analysis (TGA-DTA) and infrared (IR) methods. The results of the TEM analysis (Figure 1) show that the average size of MNP-NMe₂ and MNP-NH₂ lies in the range of 10–15 nm. The shape of the magnetite nanoparticles modified with amine groups is spherical in both cases. In addition, TEM analysis confirms that a silica shell is not constructed on the surface of the magnetite nanoparticles even though the silane monomers were added in large excess. This result is also confirmed by XRD analysis (Figure 2) that shows just characteristic peaks of magnetite nanoparticles at 2θ = 30.1°, 35.5°, 43.3°, 53.6°, 57° and 62.6°. These peaks are attributed respectively to (2 2 0), (3 1 1), (4 0 0), (4 2 2), (5 1 1) and (4 4 0) phases of a cubic unit cell.

The loading of the amine groups in the magnetically separable base catalysts was determined by TGA analysis that was performed under air atmosphere. This analysis shows that MNP-NH₂ contains about 0.46 mmol/g amine groups while MNP-NMe₂ contains about 0.63 mmol/g amine groups. Figure 3(a) shows TGA-DTA curves of the catalysts MNP-NH₂. The DTA curve contains four distinct derivative peaks, which correspond to four mass losses in the TGA curve. Between 80 °C and 118 °C there is an endothermic peak which belongs to desorption of volatile organic solvents and the percentage of the mass loss is 1.3%. At temperatures of 186 °C and 294 °C there are two exothermic peaks accompanied with mass loss of 4.5%, the first peak most likely corresponds to the decomposition of the amine groups and the second peak corresponds to the decomposition of the entire organic groups. The exothermic peak at 737 °C most probably corresponds to a phase transition in which magnetite is converted to FeO. This phase transition is accompanied by a mass loss of 1.1%. The TGA-DTA curves of MNP-NMe₂ (Figure 3(b)) show four steps of mass loss. The first mass loss of 2.7% at 80 °C takes place because of desorption of volatile organic solvents. The decomposition of the organic groups supported on the magnetite nanoparticles occurs at temperatures of 305 °C and 401 °C. In this case, the mass losses for the degradation of the organic groups are about 10.9%. According to the DTA curve for catalyst MNP-NMe₂, the phase transition of magnetite to FeO happens at a temperature of 775 °C.

The IR spectrum of catalyst MNP-NH₂ (Figure 4 (b)) shows a broad absorption band at 3430 cm⁻¹ that is attributed to N-H stretching. The peaks that appear at 1631 cm⁻¹ and 1560 cm⁻¹ belong to primary N-H bending vibration while the band at 1462 cm⁻¹ corresponds to secondary N-H bending vibration. Figure 4 (a) shows the IR spectrum of the catalyst MNP-NMe₂. In this spectrum, it can be seen that there is a broad peak at 3277 cm⁻¹, which is attributed to N-H stretching vibration of the urea group. The absorption band at 1625 cm⁻¹ belongs to C=O stretching vibration while the peak at 1555 cm⁻¹ is due to N-H bending vibration. The peak that appears at 1254 cm⁻¹ is ascribed to C-N stretching vibration.

The magnetically separable heterogeneous base catalysts were prepared using a sol-gel process. Thus, the process of preparation (Scheme 2) was based on copolymerization of either 1-(2-(dimethylamino)ethyl)-3-(3-(trimethoxysilyl)propyl)urea or N-[3-(trimethoxysilyl)propyl]-ethylenediamine with tetraethoxysilane (TEOS) in the presence of magnetite nanoparticles supported with ionic liquid groups (MNP-IL-C₄) that were prepared by reacting Fe₃O₄ nanoparticles with 1-butyl-3-(3-trimethoxysilylpropyl)-1H-imidazol-3-ium chloride in ethanol. MNP-IL-C₄ was utilized in this process as it is highly dispersible in methanol, which was used as a medium for preparing the silica sol-gel matrices. TEOS was first hydrolyzed using acid catalyst, and then the silane monomers functionalized with amine groups were added. Since the hydrolysis of the silane monomers containing amine groups is much faster than with TEOS, we decided to first hydrolyze TEOS under acidic conditions in order to bring on the end of the process to homogeneous distribution of the amine groups on the formed silica sol-gel backbone. After the copolymerization process was completed, black wet gels were obtained that were then dried to give black xerogels that could be strongly attracted to an external magnetic field. By this way of preparation, two types of magnetically separable heterogeneous base catalysts were synthesized: the first catalyst contains secondary and primary amine groups (SG-MNP-NH₂) while the second catalyst contains tertiary amine groups (SG-MNP-NMe₂).

The two heterogeneous base catalysts, SG-MNP-NH₂ and SG-MNP-NMe₂, were characterized by TEM, TGA-DTA and XRD methods. According to TEM analysis (Figure 5), it can be seen for both heterogeneous catalysts that the magnetic nanoparticles supported with ionic liquid groups (MNP-IL-C₄) are embedded in the silica sol-gel matrices. It seems that the distribution of the magnetic nanoparticles in the silica sol-gel matrices is homogeneous in the case of SG-MNP-NH₂ while it is not uniform in the case of SG-MNP-NMe₂. XRD analysis (Figure 6) confirmed the presence of magnetic nanoparticles in the silica sol-gel matrices. Thus, the XPD patterns of SG-MNP-NH₂ (Figure 6 (b)) and SG-MNP-NMe₂ (Figure 6 (a)) contain the characteristic peaks of the magnetite nanoparticles in addition to the broad peaks between 2θ = 12° and 2θ = 33° that correspond to amorphous silica produced by the sol-gel process.

In addition, TGA-DTA techniques were utilized to determine the thermal behavior of the catalysts SG-MNP-NH₂ and SG-MNP-NMe₂. The representative thermograms are shown in Figure 7. The DTA results of the catalyst SG-MNP-NH₂ (Figure 7(a)) show a broad endothermic peak centered at 125 °C which can be attributed to the removal of physical adsorbed organic volatiles and water. The mass loss after this desorption process is about 6%. In addition, the strong exothermic peak at 266 °C and the weak exothermic peak at 425 °C are accompanied by a mass loss of 12%. These peaks correspond to the combustion of the organic groups in the silica sol-gel matrix. The broad exothermic peak in a temperature range of 480–630 °C is likely to be attributed to the dehydroxylation of the silica sol-gel and the phase transition from gel to glass. These processes are accompanied by a mass loss of 7%. The DTA curve of the catalyst SG-MNP-NMe₂ (Figure 7(b)) shows a broad endothermic peak centered at 134 °C, which belongs to the desorption of organic solvents and water. The mass loss of this desorption process is 2.5%. The combustion of the organic groups of the catalyst SG-MNP-NMe₂ occurs at temperatures of 255 °C and 332 °C, and this process leads to a mass loss of 21.5%. On the other hand, the exothermic peak at 485 °C accompanied with a mass loss of 7%, can be attributed to the dehydroxylation of the silica sol-gel.

Our catalytic studies were initiated by testing the activity of the quasi-homogeneous base catalysts MNP-NH₂ and MNP-NMe₂ in a Knoevenagel reaction. Thus, malononitrile and benzaldehyde were reacted at 80 °C in the presence of 2 mol of the base catalysts dispersed in different solvents: DMSO, DMF, ethanol and water. We chose to use these polar solvents as a medium for the condensation reaction because the catalysts were more dispersible in them than in non-polar solvents like toluene, cyclohexane and heptane. After 1 h, the two catalysts gave quantitative yields of the condensation product in all of these experiments and no significant influence was observed for the type of the solvent used in these reactions. The activity of the two catalysts in the Knoevenagel condensation was similar to magnetically separable systems reported before [55,56,57,58,59].

In the second stage of our study, we investigated the reactivity of the catalysts MNP-NH₂ and MNP-NMe₂ in nitroaldol reaction. Thus, nitroethane and p-anisaldehyde were reacted at 80 °C for 16 h and in the presence of one of the two quasi-homogeneous catalysts dispersed in different solvents. In these experiments, the type of the solvent and the type of the catalyst utilized in the nitroaldol reaction could significantly affect the catalytic performance. When the reaction was performed in DMSO using MNP-NH₂ as a catalyst, only 25% of the desired product, (E)-1-methoxy-4-(2-nitroprop-1-en-1-yl)benzene, was obtained. The use of the same catalyst dispersed in ethanol led to the formation of the desired product in a yield of 68%. The condensation reaction catalyzed by MNP-NH₂ in toluene, cyclohexane and heptane did not yield any product. It is worth mentioning that the catalyst was not dispersible in these solvents and in fact worked as a heterogeneous catalytic system. It seems that the strong aggregation of the catalyst MNP-NH₂ in non-polar solvents lead to a limited accessibility to the active sites of the catalyst. When MNP-NMe₂ was utilized as a catalyst for the nitroaldol condensation in different solvents, no products were detected.

The nitroaldol reaction of nitroethane with p-anisaldehyde was also investigated using SG-MNP-NH₂ and SG-MNP-NMe₂ as magnetically separable heterogeneous base catalysts. The heterogeneous condensation reaction was optimized using different organic solvents. The results are summarized in Table 1. The results clearly show that non-polar solvents are preferable as a medium for the condensation reaction to polar solvents. The highest yield of the product (E)-1-methoxy-4-(2-nitroprop-1-en-1-yl)benzene was obtained when cyclohexane was applied as solvent. The heterogeneous catalyst SG-MNP-NMe₂ led to the formation of 1-(4-methoxyphenyl)-2-nitropropan-1-ol as main product. In this case, it was also found that non-polars solvent are preferable to polar solvent as the product was produced a higher yield.

The heterogeneous nitroaldol reaction using the catalyst SG-MNP-NH₂ was then tested in the condensation of several aromatic aldehydes with nitroethane. The condensation reactions were performed at 80 °C and cyclohexane was applied as a solvent for these reactions. The results of these reactions are presented in Table 2. The results show that the steric effect in this heterogeneous condensation reaction is negligible. Thus, when 2-methoxybenzaldehyde or 2-chlorobenzaldehyde is utilized as a substrate, high yields of the desired products are obtained. Aromatic aldehydes substituted with strong electron withdrawing groups lead to the formation of the condensation products in moderate yield. On the other hand, aromatic aldehydes containing electron-donating groups such as isopropyl give excellent yields of the condensation products. In addition, the catalyst SG-MNP-NH₂ can be recycled in the condensation of p-anisaldehyde with nitroethane for four consecutive runs without significant decrease in its activity or selectivity.