Knoevenagel condensation between a C=O group and an activated methylene group is one of the most useful carbon-carbon bond coupling reactions used to prepare several important chemical intermediates for the pharmaceutical industry. In general, organic bases or Lewis acid catalysts in a homogeneous system effectively catalyze this reaction. However, it is difficult to reuse homogeneous catalysts and their utilization leads to a number of environmental problems. The utilization of a heterogeneous catalyst instead of the homogeneous counterpart is a promising way to overcome these problems. Some important advantages of heterogeneous systems over homogeneous counterparts are their easier handing, simple workup and recyclability [1]. For this reason, several alternative heterogeneous catalysts have been successfully applied in Knoevenagel condensation, such as zeolites [1, 2], nitrogen containing carbons [3], organically functionalized mesoporous silica [4] and metal organic frameworks [5].
In the last decade, metal organic frameworks (MOFs) have attracted considerable attention for heterogeneous catalysis because of the ability to tailor the pore size and chemical functionality over a wide range [6, 7]. Although it has been well proven that nanosized porous materials exhibit better catalytic efficiency because of less diffusion limitations, their complete separation from the reaction mixture is not easy. Magnetic separation based on magnetic nanoparticles has received considerable attention and has been used in a variety of fields, mainly because of easy separation and locations. The integration of MOFs with magnetic nanoparticles could facilitate their recovery and separation. Attempts to use this strategy have been published very recently [8, 9, 10, 11, 12, 13]. Fe3O4 cored Cu3(BTC)2 magnetic microspheres have been prepared using a versatile step-by-step assembly strategy [8]. Zhang et al. [13] prepared Fe3O4 cored IRMOF-3 magnetic microspheres and tested their catalytic activities in Knoevenagel condensation reactions. The specific activity of IRMOF-3 in the microspheres was higher than the pure IRMOF-3, which having larger particles, probably because of lower diffusion of the reactants in the nanosized shell [13]. Although the Fe3O4 cored MOFs magnetic microspheres are promising alternatives for heterogeneous catalysts, reports on their preparations and applications are still limited.
The copper-based MOF Cu3(BTC)2 (HKUST-1) is a typical MOF with the above mentioned exceptional properties and has been intensively studied. This material forms face-centered cubic crystals that contain a three dimensional square-shaped channel system (1.0 nm × 1.0 nm) and accessible nanocavities. Moreover, coordinative vacancies on the Cu2+ species could be obtained by sample dehydration, which provides not only potential catalytic activity but also novel functionalized materials by surface decoration with electron rich functional groups [9, 14, 15, 16, 17]. In our initial attempt, Fe3O4 cored Cu3(BTC)2 magnetic microspheres were prepared and tested in the Knoevenagel condensation. Although it could promote the Knoevenagel condensation using the Lewis acid sites (unsaturated Cu2+ sites) and be recovered magnetically, an effort is still required to functionalize the porous frameworks with specific sites for higher efficiency. As reported, the coordinatively unsaturated Cu2+ sites could be surface decorated with electron rich functional groups to develop novel functionalized materials [9]. Recently, we published our original work on the surface modification of pure Cu3(BTC)2 with an amino-containing ligand. Although, as expected, its catalytic activity in Knoevenagel condensation as a heterogeneous catalyst was enhanced greatly, its separation from the reaction mixture through filtration was cumbersome [19]. Therefore, as schematically shown in Scheme 1, an amino-functionalized magnetic Cu3(BTC)2 with a core-shell structure was prepared and used as a magnetically separable heterogeneous catalyst for Knoevenagel condensations between benzaldehyde and ethyl cyanoacetate or ethyl acetoacetate. Importantly, this amino-containing magnetic Fe3O4@Cu3(BTC)2 not only showed high efficiency toward the Knoevenagel condensation, but could also be recovered magnetically and reused three times without obvious loss of its catalytic properties.
The powder X-ray diffraction (PXRD) patterns were recorded on Bruker D4 Endeavour with Cu Kα (λ=0.15405 nm) radiation at 40 kV and 40 mA at a scanning rate of 4°/min. Fourier transform infrared (FT-IR) spectra were recorded in the range of 400-4000 cm−1 on a Thermo Nicolet Nexus 470 Fourier transform infrared spectrometer. Scanning electron microscopy (SEM) studies were carried out with a Quanta 3D FEG instrument. Transmission electron microscopy (TEM) studies were carried out with a Tecnai G2 F20 equipped with an LaB6 filament at an acceleration voltage of 200 kV. A carbon-coated copper grid was used to prepare the samples. The Brunauer-Emmett-Teller (BET) surface area and pore volume were calculated from the nitrogen sorption isotherms measured at -196℃ on an ASAP 2000 system in the static measurement mode. The samples were outgassed at 120℃ for 3 h prior to the measurement.
The Fe3O4 nanospheres were synthesized by a solvothermal method as reported previously [20]. In a typical preparation, a solution containing FeCl3·6H2O (2.7 g, 10 mmol), NaAc (5.7 g, 70 mmol) and 50 mL of ethylene glycol was stirred for 1 h at 50℃, and then the mixture was placed in an autoclave and heated to 200℃ for 8 h. After cooling, the black magnetic microspheres were magnetically collected, followed by washing with ethanol. The solid was dried at 60℃ in an oven and denoted as Fe3O4.
Then, 3 g of as-prepared Fe3O4 was dispersed into 240 mL of an ethanol solution of mercaptoacetic acid (1.74 mmol/L). The product was magnetically collected and washed with distilled water and ethanol. The obtained solid (denoted as MAA-Fe3O4) was re-dispersed in ethanol. The process for the preparation of Fe3O4@[Cu3(BTC)2]n is schematically shown in Scheme 1. First, 3.0 g of MAA-Fe3O4 was dispersed in 240 mL of Cu(CH3COO)2·H2O ethanol solution (10 mmol/L). After stirring for 30 min at 70℃, the solid was magnetically separated from the solution and re-dispersed in 240 mL of benzenetricarboxylic acid (H3BTC) ethanol solution (10 mmol/L). The solid was magnetically collected again after stirring for 1 h at 70℃. After a given number of cycles, the samples were washed with ethanol and dried under vacuum at 100℃. The resulting composite materials were abbreviated as Fe3O4@[Cu3(BTC)2]n according to their cycle numbers.
Amino-functionalized Fe3O4@[Cu3(BTC)2]25-NH2 was prepared by surface grafting using 3-aminopropyltriethoxysilane (APS) as the modifier. As illustrated in Scheme 1, typically, 0.1 g of Fe3O4@[Cu3(BTC)2]25 was dehydrated at 120℃ under vacuum for 4 h and then 1 mL of APS in dry toluene (50 mL) was added. The resultant mixture was heated to reflux under N2 for 24 h, during which the amino groups were grafted onto the supports. After being cooled, magnetically separated and washed thoroughly with toluene and ethanol, the solid was dried under vacuum at 60℃ overnight to give Fe3O4@[Cu3(BTC)2]25-NH2.
The reaction was performed in a 25-mL round bottom flask with a magnetic stirrer. Typically, an amount of catalyst (0.21 mmol -NH2, 3%) and 8 mmol of benzaldehyde were mixed in 5 mL of solvent. After temperature adjustment under nitrogen, the reaction was started by adding 7 mmol of ethyl cyanoacetate. The conversion of ethyl cyanoacetate was periodically analyzed by gas chromatography (GC-2010 plus, Shimadzu) equipped with Rxi@-1 capillary column (30 m × 0.25 mm × 0.5 μm) and flame ionization detector (FID).
The reusability of the magnetic catalysts was tested. The catalyst was magnetically recovered and washed with ethanol (3 × 5 mL). Then it was reused directly without further purification for the next run with fresh benzaldehyde and ethyl cyanoacetate. It was used for three consecutive runs.
To test whether the reaction is a heterogeneous process, a hot filtration test was carried out. The reaction was conducted and monitored as described above, but the catalyst was magnetically removed after 30 min.
Magnetic Fe3O4 nanoparticles were first prepared from FeCl3·H2O with ethylene glycol as reducing agent and sodium acetate (NaAc) as stabilizer in a solvothermal reaction [20]. Fe3O4@Cu3(BTC)2 composites were synthesized using a step-wise assembly strategy described by Ke et al. [8]. The crystal nature and composition of the as-synthesized materials were first characterized by PXRD and the results are summarized in Fig. 1. The PXRD patterns of naked (Fig. 1(1)) and parceled (Fig. 1(2)-(5)) Fe3O4 nanoparticles all present characteristic peaks at 2θ=30.1°, 35.5°, 43.1°, 53.4°, 57.0° and 62.6°, corresponding to the (220), (311), (400), (422), (511) and (440) planes, respectively. This indicates that the particles were pure Fe3O4 with a spinel structure and its crystal phase remained intact during the parceling process. In particular, as shown in Fig. 1(2)-(5), the Fe3O4@[Cu3(BTC)2]n exhibit both characteristic peaks of Fe3O4 and Cu3(BTC)2, revealing the coexistence of Fe3O4 and Cu3(BTC)2 in the composite materials. It can also be observed that the peak intensities indexed to Cu3(BTC)2 increase with the wrapping steps. Furthermore, the weak and broad diffraction peaks of Cu3(BTC)2 confirmed the formation of nanosized Cu3(BTC)2 crystals on the surface of the magnetic particles, and the size was estimated to be approximately 26 nm by applying the Scherrer formula on the diffraction peaks of Cu3(BTC)2. Post-modification of Fe3O4@[Cu3(BTC)2]25 with APS under mild condition was done to widen its catalytic applications. As expected, their XRD patterns (Fig. 1(4) and (5)) were very similar before and after modification, indicating the structure of Fe3O4@[Cu3(BTC)2]25 remained intact.
Although PXRD results demonstrated that the nanocomposites were composed of Fe3O4 nanoparticles and Cu3(BTC)2, it did not confirm if the nanocomposite is a real Fe3O4@[Cu3(BTC)2]n nanocomposite or just a physical mixture of two separate phases of Fe3O4 and Cu3(BTC)2. To gain further insight into the microstructure of the as-synthesized Fe3O4@Cu3(BTC)2 nanocomposites, the materials were further characterized by SEM and TEM techniques. Representative SEM (Fig. 2(a)) and TEM (Fig. 2(d)) images reveal that the prepared Fe3O4 nanoparticles were spherical in shape and had a diameter distribution of around 100-600 nm. As expected and shown in Fig. 2(e), the TEM image of the formed nanocomposites clearly show core-shell structures with Fe3O4 as the core and Cu3(BTC)2 as the shell. The thickness of the shell is about 200 nm after 25 assembly cycles. No obvious structural change was observed after surface modification when comparing their SEM and TEM images (Fig. 2(b)-(c) and 2(e)-(f)). The amino-group content in Fe3O4@Cu3(BTC)2-NH2 calculated from SEM-EDS results is approximately 2.1 mmol/g (Fig. 3). It can be concluded that Cu3(BTC)2 was successfully wrapped on the surface of the Fe3O4 nanoparticles and the amount of Cu3(BTC)2 could be controlled by the number of wrapping cycles.
Fig. 4 shows the FT-IR spectra of the bare Fe3O4 and Fe3O4@[Cu3(BTC)2]25, as well as the amino-functionalized sample. Typical peaks at 580 cm−1 for naked Fe3O4 originated from the symmetric stretching of Fe-O-Fe. Features at 1620, 1560 and 1720 cm−1 are ascribed to the benzene skeleton vibration and the stretching vibration of the C=O of the ligands in Fe3O4@[Cu3(BTC)2]25. After the introduction of amino groups, typical peaks at 3440 and 3490 cm−1 corresponding to the symmetric and asymmetric stretching of primary amines as well as typical peaks at 2920 and 2845 cm−1 associated with C-H vibrations were observed. This results confirm the introduction of amino-containing ligands onto Fe3O4@[Cu3(BTC)2]25. The amino group should be attached to the Fe3O4@Cu3(BTC)2 through a coordination mode. Both amino and methoxy groups in the APS molecule competitively play a role as possible coordination sites to the coordinatively unsaturated Cu2+ sites.
The N2 adsorption-desorption experiments were used to verify the porous structure of the hybrid materials as well as to calculate their surface area and pore volumes. As shown in Fig. 5, the completely reversible isotherm of Fe3O4@[Cu3(BTC)2]25 exhibited Type-I behavior, typical for microporous materials. For comparison, the BET surface area and pore volumes are summarized in Table 1. The BET surface area of Fe3O4@[Cu3(BTC)2]25 is 462 m2/g and the total pore volume is 0.38 cm3/g, which were comparable to those reported by Ke et al. [8]. The amino-modified sample Fe3O4@[Cu3(BTC)2]25-NH2 exhibited a significant decrease in both surface area and pore volume. The BET surface area decreased significantly from 462 m2/g to 129 m2/g, and the total pore volume decreased from 0.38 cm3/g to 0.16 cm3/g, after amino-ligand grafting. Obviously, the decreases are caused by the inclusion of APS into the nanocages. Similar results were also observed for the surface modification of pure Cu3(BTC)2 in our original work [19].
Knoevenagel condensation has been employed as a classic test reaction to analyze the activity of a variety of solid base catalysts. Therefore, the Knoevenagel condensation between benzaldehyde and ethyl cyanoacetate was selected as a model reaction to evaluate the catalytic activity of Fe3O4@[Cu3(BTC)2]25-NH2. Solvent polarity has strong effects on the catalytic Knoevenagel reactions [18, 21, 22, 23, 24]. To establish the solvent effects for Fe3O4@[Cu3(BTC)2]25-NH2, the reaction was carried out in a variety of organic solvents and the results are summarized in Table 2. The highest activity of Fe3O4@[Cu3(BTC)2]25-NH2 was found in DMSO, which had the highest dielectric constant (Ɛ=48.9). Conversion of ethyl cyanoacetate in dimethyl sulfoxide (DMSO) was 97% after 1 h (Table 2, Run 6). Ethanol (a protic solvent, Ɛ=24.3) was also found to be a good medium and the conversion of ethyl cyanoacetate was 84% after 1 h (Table 2, Run 7) under the same conditions. By contrast, much lower activities were found in tetrahydrofuran (THF) (Ɛ=7.5), toluene (Ɛ=2.4) and hexane (Ɛ=1.9), indicating that the aprotic and less polar solvents are not good media for this reaction. For example, conversion of ethyl cyanoacetate was only 8% after 1 h in the presence of hexane (Table 2, Run 8). The results above are consistent with the reported conclusions that the polarity of the media has a strong effect on the reaction route and reaction rate of the Knoevenagel condensation reaction [18, 21, 22, 23, 24]. As reported, the reaction rate for the IRMOF-3 catalyzed Knoevenagel condensation between benzaldehyde and ethyl cyanoacetate was much higher in polar solvents than in nonpolar solvents [18]. A similar effect was also observed for the amino-containing silica catalyzed reaction [4]. The strong difference in activity is probably because of the polarity and amphiprotic properties of the solvents. The polar solvents could stabilize the carbanion intermediate and their amphiprotic properties could favor the electrophilic polarization of the carbonyl group in benzaldehyde [22]. All results above demonstrated that amino-tagged MOFs performed better in highly polar solvents or protic solvents.
Apart from good catalytic activity and selectivity, long-term stability is an important criterion for solid catalysts. As mentioned above, the best activity was found in DMSO. Therefore, to check the stability for Fe3O4@[Cu3(BTC)2]25-NH2, the catalyst was recycled for the Knoevenagel condescension between benzaldehyde and ethyl cyanoacetate in DMSO (Table 3). After each cycle, the solid was magnetically separated, washed with DMSO and then reused in the next cycle. The conversion in the second run (Table 3, Run 2) was much lower than the first run (Table 3, Run 1), and the PXRD patterns (Fig. 7(3)) of the recovered catalyst demonstrate that its crystal structure was completely destroyed by the reaction process, with only a broad peak evident between 5°-30°. Its SEM image (Fig. 8(b)) also confirms the disappearance of the shell, probably resulting from metal-DMSO coordination. Ethanol was chosen as an alternative solvent for recyclability (Table 3, Run 3-5). It is noteworthy that, as we expected, the catalyst activity is stable during the recycle reaction, and the stability of the structure was further confirmed by PXRD (Fig. 7(2)), SEM (Fig. 8(c)) and TEM(Fig. 8(d)). The catalyst could be recovered magnetically from the reaction medium and recycled three times without obvious loss of the activity and selectivity. The minor loss of the activity is probably owing to the loss of solid during the separation, rather than from the leaching of the amino groups from the support. Thus, a sacrificial balance between activity and recyclability needs to be made.
In this work, a novel core-shell structured Fe3O4@[Cu3(BTC)2]25-NH2 material was successfully fabricated as active solid base catalysts for Knoevenagel condensation. Core-shell structure was confirmed by PXRD and TEM techniques, its composition was confirmed by SEM-EDS and the amine functionality was confirmed using FT-IR spectroscopy. The core-shell Fe3O4@[Cu3(BTC)2]25-NH2 material is ideal recyclable catalysts for Knoevenagel condensation by combining a superparamagnetic core for efficient magnetic separation and a porous MOF shell for high catalytic activity. The remarkable activity for Knoevenagel condensation reactions is ascribed to the presence of amino groups as well as Lewis acid copper ions in Fe3O4@[Cu3(BTC)2]25-NH2. Obvious solvent effects on the conversion and stability were observed and ethanol was selected as a good medium providing a balance between activity and recyclability. The catalyst could be easily separated from the reaction mixture and reused without significant change in the catalytic activity after three reaction cycles.
2016, Vol. 37 Issue (3): 420-427
PDF (1115 KB)
2016, Vol. 37 Issue (3): 420-427
PDF (1115 KB)