Development of basicity in mesoporous silicas and metallosilicates

D. Kryszak ab, K. Stawicka a, M. Trejda a, V. Calvino-Casilda *b, R. Martin-Aranda b and M. Ziolek *a
aFaculty of Chemistry, Adam Mickiewicz University in Poznan, Umultowska 89b, 61-614 Poznan, Poland. E-mail: ziolek@amu.edu.pl
bDepartamento de Química Inorgánica y Química Técnica, UNED, Senda del Rey, 9, E-28040-Madrid, Spain. E-mail: vcalvino@ccia.uned.es

Received 9th May 2017 , Accepted 9th June 2017

First published on 9th June 2017


We report herein an experimental study on the development of basicity on mesoporous silicas and metallosilicates (Nb- and Ce-) on SBA-15 and MCF porous structures. Our results demonstrated that the role of the catalyst structure (hexagonally ordered mesoporous SBA-15 vs. mesoporous cellular foam MCF), the effect of the amount and nature of metal in the supports (different acidic/redox properties), and the effect of the type of nitrogen-containing organic modifier (imidazole, triazole and aminotriazole) play an important role on the activity and selectivity in Knoevenagel condensation, which is a well-known probe reaction to determine the basicity of inorganic solids. We demonstrate that niobium Lewis acidity is determinant for the Knoevenagel condensation and, also, redox properties of cerium play an important role in the basicity. In fact, acid–base cooperation plays a positive role in promoting the Knoevenagel condensation via the ion-pair mechanism. These mesoporous silicas and metallosilicates are interesting catalysts for production of fine chemicals due to their thermal stability and controllable acid–base catalytic properties.


1. Introduction

Many catalysts active in the production of fine chemicals and pharmaceuticals require the presence of basic centers which may be introduced to solid samples using an impregnation or grafting strategy. Among the reactions involving basic species are Knoevenagel1 or Claisen–Schmidt condensation,2 isomerization,3 Michael addition,4 alkylation,5 transesterification,6 and the Friedländer reaction.7–9 So far, a variety of basic heterogeneous catalysts have been reported10 as for instance materials modified with alkali metals,6,11 metal oxides2,6,12 or silanes possessing a basic nitrogen center.1,13 Basic active centers loaded on supports can be accompanied with additional species which enhance the catalytic activity of basic samples.11,14 It is well known that nucleophilic oxygen or nitrogen in the surroundings of different atoms can be a source of basicity.14–16 Usually basic species are loaded on different supports. Requirements for a heterogeneous catalyst to be used as the support for basic centers are met by mesoporous silica like SBA-15 and MCF. They exhibit high thermal stability, very high surface areas allowing the loading of high amounts of the active phase and an ordered system of large pores or cavities facilitating the diffusion of reagents. The problem which appears by using this type of support is its interaction with strong alkali solutions, usually used for generation of basic centers, leading to the destruction of the support material. This problem can be avoided by a multi-coating strategy,17,18 a redox strategy19,20 or application of organic compounds containing nitrogen, playing the role of basic centers used as modifiers loaded in mesoporous materials. The redox strategy implies the formation of a carbon coat on the surface of walls in e.g. SBA-15 silica by temperature treatment of the template containing mesoporous silica in the presence of nitrogen and, next, loading of alkali nitrate. The reducibility of the carbon layer causes the lowering of the temperature (400 °C instead of 600 °C used in the classical method) necessary for transfer of nitrate into alkali metal oxide and therefore protects mesoporous silica from destruction of its ordered structure. A similar effect can be reached by coating of pores with zirconia.17 The functionalization of mesoporous silica by organic compounds containing nitrogen as a source of basicity also allows the protection of the support structure and gives rise to active catalysts. For example, amino-grafted mesoporous cellular foams, MCF, were applied in quinoline synthesis and Friedländer condensation.21 Different ordered mesoporous silicas and metallosilicates were used as supports for different aminoorganosilanes (APMS) and applied as effective catalysts in Knoevenagel condensation.14,16,22,23

The appropriate choice of the active phase adapted to the catalytic process planned to be performed is a key point for the design of effective catalysts. There is no doubt that the neighbors of basic centers determine their strength and thus their activity. Therefore, the use of modifiers containing more than one nitrogen atom, such as imidazole, triazole or 1,2,3-triazol-4-ylmethanamine, can lead to obtaining appropriate basic properties addressed to the desired reaction. Moreover, there are some base catalyzed reactions (e.g. Knoevenagel condensation) which exhibit higher activity if in addition to the basic centers the catalysts contain acidic sites (mainly Lewis acid sites, LAS).14,24 The required acidic sites can be located in the support.

In our recent paper24 we have immobilized imidazole in mesoporous SBA-15 and MCF silicas and niobosilicates and tested them in Knoevenagel condensation. It was found that the catalytic activity of imidazole species was enhanced by the presence of niobium in the support, working as LAS and taking part in the reaction. In this work, we extend the previous study by comparing the role of niobium (in different loadings and incorporated by post synthesis modification) with the effect of cerium species on the activity of imidazole anchored. The choice of ceria as a modifier was dictated by its unique properties, different than that of the niobia species. The existence of vacancies on the surface of CeO2, which is beneficial for the generation of basicity,10 makes it an interesting candidate for modification of the silica support towards achievement of the dual effect, which enhances basicity of nitrogen-containing organic compounds. Moreover, in this work, the other nitrogen-containing organic compounds, triazole and 1,2,3-triazol-4-ylmethanamine, anchored into SBA-15 and MCF supports were prepared and characterized. Their activity was tested in the Knoevenagel condensation performed using different carbonyl compounds with different pKa values. The goal of this work was to consider the following features: i) the role of the catalyst structure (hexagonally ordered mesoporous SBA-15 vs. mesoporous cellular foam MCF), ii) the effect of the metal nature and amount in the supports (different acidic/redox properties), and iii) the effect of the type of nitrogen-containing organic modifier on the activity and selectivity in the Knoevenagel condensations.

2. Experimental

2.1. Synthesis of catalysts

The synthesis procedures of MCF and SBA-15 supports were performed according to ref. 24. This route is described in detail in the ESI (S1).

The modification of MCF and SBA-15 with niobium or cerium species was performed using the impregnation method as follows. At the beginning, the metal precursor, i.e. ammonium niobate(V) oxalate hydrate (Aldrich, 99.99%) or cerium(III) nitrate hexahydrate (Aldrich, 99.99%), was first dissolved in distilled water. The amount of the metal precursor was chosen to achieve a Si/metal molar ratio of 19 or 103. The obtained salt's solution was then added to dried silica (dried at 100 °C for 24 h), which was previously evaporated for 1 h at 80 °C. The obtained mixture was then stirred in a vacuum evaporator at 80 °C until water removal. Then the solid material was dried at 120 °C for 12 h and calcined at 550 °C for 5 h with a temperature ramp of 5 °C min−1.

The obtained MCF, SBA-15, Nb/MCF, Nb/SBA-15, Ce/MCF and Ce/SBA-15 with different metal loadings (Si/Nb or Ce = 19 and 103) were then modified with imidazole followed by (3-chloropropyl)trimethoxysilane (ClPTMS) modification as in ref. 24. The detailed procedure is described in the ESI (S2).

For triazole modified materials, almost the same modification procedure was used as for the imidazole functionalized samples. The only difference was adding 1.3 cm3 of 1H-1,2,3-triazole (Aldrich) instead of imidazole to the toluene solution.

The synthesis procedure of 1,2,3-triazol-4-ylmethanamine modified SBA-15 and MCF is described in ref. 25.

2.2. Characterization of catalysts

The obtained samples were characterized by N2 adsorption/desorption, XRD, elemental analysis, FT-IR with pyridine adsorption as a probe molecule, DTA/TG and UV-vis DR. These techniques are precisely described in ref. 14 and 24, and the details are presented in the ESI (S3).

2.3. Knoevenagel condensations

The catalysts were examined in Knoevenagel condensation between benzaldehyde and ethyl cyanoacetate (pKa = 9), ethyl acetoacetate (pKa = 10.7) and diethyl malonate (pKa = 13.3). These reactions permitted assessment of the basic strength of the tested samples. It is well known that the abstraction of a proton from a methylene compound exhibiting higher value of pKa required the presence of a catalyst possessing stronger basic sites. The reaction conditions of Knoevenagel condensations are precisely described in ref. 14 and shown in the ESI (S4).

3. Results

The idea of this work was to anchor different nitrogen-containing organic compounds into SBA-15 and MCF structured silica and metallosilicates, as well as to characterize their surface properties and catalytic activity in Knoevenagel condensation reactions. For this purpose, imidazole and triazole were immobilized onto SBA-15, Nb/SBA-15, Ce/SBA-15, MCF, Nb/MCF, and Ce/MCF mesoporous supports via functionalization of the supports with (3-chloropropyl) trimethoxysilane (ClPTMS) and anchoring of the nitrogen-containing compounds in their interaction with chloride ions. Concerning 1,2,3-triazol-4-ylmethanamine, it was introduced via the click reaction between propargylamine and (3-azidopropyl)(triethoxy)silane loaded on the supports.25 The way of the anchoring of the active species is illustrated in Scheme 1.
image file: c7cy00927e-s1.tif
Scheme 1 The way of anchoring of imidazole, triazole and aminotriazole active species.

3.1. Characterization of catalysts

3.1.1. Structural/textural properties of the supports. Two different structures (SBA-15 and MCF) and three different compositions (silica, silica modified with niobium and silica modified with cerium species) of the supports for the nitrogen-containing organic compounds were applied in this work. The XRD patterns (Fig. 1) clearly indicate that the SBA-15 structure was preserved after modification with niobium and cerium species. The patterns of all SBA-15-based supports exhibited one main peak at 2 theta ca. 1° (100) and two small, well-resolved peaks between 1.5° and 2° ((110) and (200)). All peaks are characteristic of hexagonally ordered mesopores in the SBA-15 structure. The presence of two small peaks ((110) and (200)) indicates long-range ordering. In the case of the niobium-containing material, the position of the main peak is slightly shifted because of the change in the unit cell caused by metal incorporation into the silica skeleton.
image file: c7cy00927e-f1.tif
Fig. 1 Small and wide angle XRD patterns of the supports and CeO2. (CeO2 - JCPDS ICDD PDF Card-00-043-1002).

The XRD patterns of SBA-15 and MCF materials modified with cerium show the reflections typical of the cubic CeO2 structure (JCPDS ICDD PDF Card – 00-043-1002). They are clearly seen if Si/Ce = 19 and at the lower metal loading (Si/Ce = 103), they are poorly visible. In contrast, the XRD patterns of all niobium-containing supports did not reveal any niobium oxide crystalline phase. There are two possible explanations for this fact: i) the niobium oxide is so well-dispersed that the crystallites are not detectable by the XRD technique, or ii) niobium is not present in the form of crystalline extra-lattice metal oxide but is included on the surface of silica or is present in the amorphous form of the metal oxide. More information about the metal species comes from the UV-vis DR spectroscopy study. Fig. 2 presents the UV-vis DR spectra of the samples with the higher metal loading. The spectra of niobium-containing SBA-15 and MCF showed two bands characteristic of charge transfer between oxygen ions and niobium (+5) cations in tetrahedral coordination (223 nm) and penta-coordination (261 nm or 251 nm), both typical of niobium located in the silica matrix.24,26–28 Moreover, the lack of a band at ca. 330 nm, typical of bulk Nb2O5 (ref. 26), was consistent with the XRD results and indicated that niobium is located on the silica matrix and not in the extra-framework position. The spectra of cerium containing samples showed one main band at ca. 290 nm and a smaller one at ca. 220 nm, both characteristic of bulk crystalline cerium oxide.29–31 Thus, the UV-vis and XRD study allowed us to conclude that niobium was located on the mesoporous silica matrix, whereas cerium species were included in the extra-framework crystalline oxides, irrespective of the structure of mesoporous silica (SBA-15 or MCF).


image file: c7cy00927e-f2.tif
Fig. 2 UV-vis DR spectra of the supports containing a higher amount of metals and the samples after imidazole immobilization.

Table S1-SD shows the textural parameters calculated from nitrogen adsorption/desorption isotherms of SBA-15 and MCF materials also modified with metal species, i.e. cerium or niobium. These parameters changed for both mesoporous silicas after modification. In particular, the total surface area, micropore area as well as the pore volume and diameter, decreased after incorporation of Nb or Ce. The differences in textural parameters between the supports before and after modification were more pronounced for higher metal loading, independent of the kind of metal used.

3.1.2. Acidity of the supports. Acidity of the supports was investigated by pyridine adsorption combined with an FTIR study. Fig. 3 presents the spectra obtained after pyridine adsorption at 150 °C and desorption at the same temperature and at 200 °C. Pyridine adsorption on cerium-containing MCF supports led to the appearance of IR bands at 1446 cm−1 and 1596–1597 cm−1 typical of vibration of pyridine hydrogen bonded to the solid surface.32–34 Interestingly, the intensity of these bands was much lower in the spectrum of the sample containing a higher amount of cerium (Si/Ce = 19). This phenomenon suggests that cerium species did not participate in the chemisorption of pyridine. This suggestion was confirmed by the analysis of the FTIR spectra in the hydroxyl region. It is clearly demonstrated that silanol groups of the support (the band at 3745 cm−1) took part in the chemisorption of pyridine. This interaction was shown to be stronger in Ce/MCF-103 than in Ce/MCF-19 because a higher loading of cerium species diminished the number of silanol groups available for the interaction with pyridine. After pyridine adsorption, a significant decrease in the intensity of the band at 3745 cm−1 was accompanied by the appearance of a broad band between 3700 and 3650 cm−1 typical of hydrogen bonded silanol groups. This broad band disappeared after desorption of pyridine. Pyridine desorption caused also the rebuilding of silanol groups leading to the same intensity of the IR band as that after activation (evacuation at 350 °C) of the sample. Finally, the FTIR spectra after pyridine adsorption and desorption on cerium modified MCF are the same as that obtained on pristine MCF mesoporous silica. The presented results indicate that cerium species did not play the role of acidic centers. Weak acidity was due to the presence of silanol groups in the support.
image file: c7cy00927e-f3.tif
Fig. 3 FTIR spectra of the supports after evacuation at 350 °C (a) followed by pyridine adsorption at 150 °C (b), desorption at 150 °C (c), 200 °C (d), 250 °C (e) and 300 °C (f). In the spectra in the range of 1700–1400 cm−1, the spectrum after activation was subtracted.

In contrast to Ce/MCF materials, the niobium-containing samples revealed the presence of vibrations in pyridine molecules coordinatively bonded to Lewis acid sites (Fig. 3; FTIR bands at 1447 cm−1 and 1610 cm−1).32,35 The first band overlapped with the band coming from vibrations in hydrogen bonded pyridine, but the band at 1597 cm−1 proved the presence of such species, similar to the cerium containing samples. Additional evidence for pyridine chemisorbed on LAS was the IR band at 1490 cm−1, typical of vibrations in both types of pyridine chemisorbed, on LAS and in pyridinium ions formed in the interaction with Brønsted acid sites (BAS). As the band at ca. 1550 cm−1 characteristic of vibration of pyridinium ions was not observed in the FTIR spectra, the one at 1490 cm−1 characterized the vibrations of pyridine chemisorbed on LAS. It is not surprising that with increasing niobium loading (from Si/Nb = 103 to Si/Nb = 19), the intensity of the infrared bands characteristic of pyridine chemisorbed on LAS increased, because LAS originate from cationic niobium species. Moreover, such chemisorbed pyridine was stable under evacuation at increasing temperatures showing the presence of relatively strong LAS on the surface of the Nb/MCF-19 support. This behavior is in line with our earlier study indicating the formation of LAS related to niobium incorporation into different silica mesoporous materials and is discussed in a review paper.36 Taking into account the extinction coefficient from ref. 33 characteristic of the IR band at 1447 cm−1 (adsorption of pyridine on LAS), the number of LAS on the surface was calculated after pyridine adsorption at 150 °C and desorption at two temperatures, 150 °C and 200 °C. The above results prove a higher amount of LAS in the samples with the higher niobium loading (Si/Nb = 19; LAS number 16.5 × 1018 after evacuation at 150 °C) than in those with Si/Nb = 103 (LAS number 9.8 × 1018 after evacuation at 150 °C), as well as similar acid strength on both Nb/MCF-19 and Nb/MCF-103 concluded from the drop in the number of pyridine molecules adsorbed on LAS after evacuation at 150 °C and 200 °C (47% drop for Nb/MCF-103 and 48% for Nb/MCF-19).

3.1.3. Composition and structural properties of the catalysts. The data from nitrogen adsorption/desorption isotherms obtained for silicas and metallosilicates modified with imidazole, triazole or aminotriazole are presented in Table S1-SD. It is clear that the surface area, pore volume and diameter decreased after anchoring of organic compounds onto the supports. This decrease was more significant for metallosilicates, which exhibited a higher metal loading. The changes in the textural/structural parameters after anchoring of the nitrogen-containing compounds certify the success of the inclusion of modifiers into the obtained materials. Estimation of the textural changes depending on the metal (cerium or niobium) loading in the support seems to be quite interesting. The N2 adsorption/desorption isotherms (Fig. S1-SD) were typical of mesoporous solids and belong to type IV according to IUPAC classification.37 The isotherms show an H2-type hysteresis at a relatively high pressure, depending on the architecture of mesoporous silica (SBA-15 or MCF) which is characteristic of mesoporous materials. This hysteresis is due to capillary condensation and evaporation from mesopores. Interestingly, loading with cerium led to a wider hysteresis loop than niobium incorporation in the MCF material because of the location of CeO2 in the pores, in contrast to the location of niobium in the silica skeleton. In the SBA-15 structure, a higher cerium loading (Si/Ce = 19) initiated the two-step desorption branch caused by partial blockage of pores by cerium oxide. Modification of the supports with imidazole or triazole did not lead to the changes in shapes of nitrogen adsorption isotherms, but decreased the volume of nitrogen adsorbed. It is important that the shapes of the pore size distribution profiles (Fig. S2-SD) in SBA-15 materials and the window size distribution in the MCF samples were not influenced by imidazole or triazole anchoring. In contrast, the PSD obtained from the adsorption branch (characteristic for cells) in MCF materials containing cerium species were different after imidazole or triazole inclusion, which suggested some changes in the cerium oxide location resulting from modification with the amines.

The anchoring of nitrogen-containing organic compounds on the supports described in the previous sections was also confirmed by the elemental analysis of the nitrogen content, the presence of vibrations in functional groups in the FTIR spectra and the characteristic bands in the UV-vis DR spectra. Table 1 presents the nitrogen content expressed as the wt% and mmol g−1 of nitrogen as well as mmol of molecules containing the imidazole or triazole ring. The number of nitrogen atoms in each modifier is different: two in imidazole, three in triazole and four in aminotriazole species formed by the click reaction. Theoretically, each nitrogen atom can take part in catalytic reactions as an active center. Therefore, the amount of nitrogen atoms expressed as wt% or mmol is important. On the other hand, if the effectiveness of modifiers' anchoring is considered, the number of modifier molecules has to be taken into account. The results collated in Table 1 show that the difference between the mmol of imidazole and triazole species included into the desired support by the same method was insignificant, whereas the amount of aminotriazole immobilized by the click reaction was much lower. The type of the support structure (bigger cells in MCF vs. smaller channels in SBA-15) had an impact on the effectiveness of modifier loading but the nature of metal species seemed to be more important. It was evidenced that for the same structure (SBA-15 or MCF), the presence of niobium enhanced the effectiveness of all modifiers' immobilization. In our earlier papers, the chemical interaction between niobium species and amine groups was proved.14,38 A similar interaction can occur with imidazole and triazole rings and it can enhance the incorporation of modifiers. The effectiveness of modifier loading is to a certain extent conditioned by the amount of the organosilane (ClPTMS) anchored in the first step of modification. ClPTMS loading was affected by metal dopants and the structure of mesoporous silica. For the MCF open structure, both metal dopants enhanced the amount of ClPTMS loaded, whereas in the case of SBA-15, the steric effect due to metal incorporation seemed to be more important and decreased the amount of ClPTMS anchored.

Table 1 Nitrogen content (from elemental analysis)
Catalyst Nitrogen [wt%] Nitrogen atoms [mmol g−1] Nitrogen compound [mmol g−1]
Ce/SBA-15-103-Im 3.21 2.29 1.15
Ce/SBA-15-19-Im 2.54 1.81 0.91
Ce/MCF-103-Im 4.03 2.88 1.44
Ce/MCF-19-Im 4.28 3.06 1.53
Nb/SBA-15-103-Im 3.54 2.53 1.26
Nb/SBA-15-19-Im 3.28 2.34 1.17
Nb/MCF-103-Im 4.45 3.18 1.59
Nb/MCF-19-Im 4.65 3.32 1.66
MCF-Tr 5.39 3.85 1.28
Ce/MCF-103-Tr 5.11 3.65 1.22
Ce/SBA-15-103-Tr 4.71 3.36 1.12
Nb/MCF-103-Tr 6.68 4.77 1.59
SBA-15-Tr-NH2 4.21 3.01 0.75
NbSBA-15-Tr-NH2 5.11 3.65 0.91
MCF-Tr-NH2 4.60 3.29 0.82
NbMCF-Tr-NH2 5.29 3.78 0.94


The successful loading of modifiers was confirmed by the FTIR spectra (Fig. 4) in which the presence of the bands from C–N (at 1565 cm−1) and C[double bond, length as m-dash]N (at 1513 cm−1) vibrations confirmed the anchoring of the imidazole ring. For the MCF-Tr sample, a broad band at 1650 cm−1 was detected which may be related to C–N (ref. 39) and C[double bond, length as m-dash]C (ref. 40) vibration in the triazole ring. Aminotriazole species exhibited additionally the bands characteristic of vibrations in amine groups (at 1419 and 1608 cm−1).40


image file: c7cy00927e-f4.tif
Fig. 4 FTIR spectra in the C–N and N–H vibration range of SBA-15 and MCF-based materials activated under vacuum for 2 h at 100 °C.

The presence of imidazole, triazole and aminotriazole was also supported by the UV-vis DR spectra (Fig. 2 and Fig. S3-SD). The band at ca. 220 nm, typical of charge transfer in imidazole and triazole rings, was particularly very visible in the samples containing smaller amounts of metals in the supports (Si/metal = 103), because the interaction between nitrogen compounds and metals in the support disturbed charge transfer in imidazole and triazole rings. For the aminotriazole modified samples, the band related to the electron charge transfer in the triazole ring was red shifted in comparison to its position in the spectra of triazole modified materials.25 This feature was caused by the interaction of amine species (the substituent acting as a donor of electrons41) with the five-membered ring of triazole.

3.1.4. Stability of nitrogen compounds used as modifiers. The stability of the active phase is an important feature for the materials addressed to catalytic processes. This feature was estimated by thermogravimetric investigations. The results for the selected samples are shown in Fig. S4-SD. DTA curves exhibited two peaks accompanied by mass loss. In our earlier paper,24 both peaks were assigned to the decomposition and oxidation of imidazole located at different sites. The first came from decomposition and oxidation of organic species incorporated into the samples modified with imidazole and anchored via a propyl chain. It was the main most intense peak on the DTA curve and corresponded to the greatest mass loss in the range of 381–400 °C. The second one, less intense in the range of 449–525 °C, originated from the decomposition and oxidation of imidazole anchored directly on siliceous atoms.24 The temperature of the peak maximum was an indicator of the modifier stability. Concerning the imidazole species, one can observe the effect of the nature of the metal in the support. This modifier was more stable on the samples containing niobium (temperature of the maximum of the main peak was 400 °C and 398 °C, depending on the support structure), than on the materials with cerium species (maximum temperature – 381 °C and 382 °C). Interestingly, the modifier was less stable if anchored on the supports containing cerium species than those with imidazole loaded on MCF (maximum temperature – 397 °C from ref. 24) and SBA-15 (maximum temperature – 389 °C from ref. 24) without any metals.

Fig. S4-SD shows also that the stability of triazole did not depend on the presence of metals in the support but the decomposition of triazole connected directly to siliceous atoms was strongly affected by the nature of metals in the supports (457 °C for Ce/MCF-103-Tr and 525 °C for Nb/MCF-103-Tr).

The aminotriazole species were the worst stabilized among all the used nitrogen-containing modifiers. This organic compound decomposed at temperatures between 320–360 °C. However, the positive impact of the MCF structure and niobium dopant, improving the stability of the aminotriazole species, was notable.25

3.2. Knoevenagel condensations

The Knoevenagel condensation between benzaldehyde and the three active methylene compounds, ethyl cyanoacetate (pKa = 9); ethyl acetoacetate (pKa = 10.7) and diethyl malonate (pKa = 13.3) was performed over imidazole immobilized on SBA-15, Nb/SBA-15, Ce/SBA-15, MCF, Nb/MCF, and Ce/MCF supports. The role of the structure of the support (SBA-15 vs. MCF) and the nature of the metal modifier in the support (Nb or Ce) was considered on the basis of the results obtained for these three condensation processes. Moreover, the condensation between benzaldehyde and ethyl cyanoacetate (pKa = 9) was carried out on the materials modified by triazole and aminotriazole (1,2,3-triazol-4-ylmethanamine). This study allowed us to determine the role of the type of nitrogen-containing organic modifier on activity and selectivity in Knoevenagel condensations. The activity and selectivity obtained in each reaction are presented in Tables S2–S4 SD.
3.2.1. The role of the support structure. Two different structures of ordered mesoporous silica and metallosilicates were applied, i.e. hexagonally ordered SBA-15 molecular sieves and mesoporous cellular foams, MCF. The structures differed not only in the size of pores indicated in Table S1-SD, but also in the type of pore system. SBA-15 materials have parallel arranged and not interconnected mesopores. Moreover, besides mesopores, they also contain micropores. MCF materials exhibit big cells and smaller windows (but still bigger than the pore diameters in SBA-15), as well as interconnections between pores. These differences in the materials used as supports for anchoring of basic modifiers could determine not only the diffusion effects but also the possibility of formation of bulky intermediates in the Knoevenagel condensation. The role of the mesoporous silica structure has been recently pointed out by Kuśtrowski et al.42 who studied the effect of the structure and architecture of SBA-16, SBA-15 and MCF on poly(vinylamine) (PVAm) loading and activity in Knoevenagel condensation. The highest amounts of the polymer were deposited on the MCF support characterized by the presence of the widest mesopores. This feature affected the activity. Thus, we could also expect higher activity of catalysts based on MCF supports.

Fig. 5 displays the diagrams of benzaldehyde conversion in the three Knoevenagel condensations performed on the materials modified with imidazole. As expected,14 their activity generally decreased with increasing pKa of the methylene compound, being the highest for pKa = 9 and the lowest for pKa = 13.3. There was some perturbation in this general order, i.e. for Nb/SBA-15-103-Im, the activity at pKa = 10.7 was lower than that at pKa = 13.3. The same tendency, but less pronounced, was observed for Ce/SBA-15-103-Im.


image file: c7cy00927e-f5.tif
Fig. 5 The role of the catalyst structure (SBA-15 vs. MCF) on the conversion of benzaldehyde in the Knoevenagel condensation performed at different pKa. Conversion of benzaldehyde after 240 min for pKa = 9; 240 min for pKa = 10.7; and 300 min for pKa = 13.3.

The effect of the structure is illustrated in Fig. 5 for lower metal loadings in the supports (Si/metal = 103). The type of structure was not important in the condensation between benzaldehyde and ethyl cyanoacetate (pKa = 9) as well as benzaldehyde and diethyl malonate (pKa = 13.3), irrespective of the nature of metal species. In contrast, more than 20% difference was noted in the conversion of benzaldehyde with ethyl acetoacetate (pKa = 10.7) performed on the metallosilicates of different structures. The catalysts based on MCF (the structure containing interconnected large cells) exhibited much higher activity than the catalysts based on SBA-15 materials. This effect was the same for niobium and cerium containing supports.

Looking for the reason why only for the reaction at pKa = 10.7 the structure of the catalysts influenced their activity, an insight into the selectivity of the reaction was made. The results are presented and discussed in the next section.

3.2.2. The effect of the metal nature in the supports. Two different metals were used for modification of the mesoporous supports, niobium and cerium. Both were introduced after the synthesis of SBA-15 and MCF silicas. However, different locations of both metallic species were achieved. Niobium was incorporated into the silica matrix, whereas cerium formed crystalline cerium(IV) oxide. It implied different impacts on the activity of imidazole anchored catalysts. The role of the metal nature and loading on activity in the Knoevenagel condensations depended on the methylenic compound used as a reagent for condensation with benzaldehyde. Therefore, the impact of the metal will be considered separately for each pKa of the methylenic compound.

At pKa = 9, metal incorporation into the supports caused some changes in benzaldehyde conversion depending on the metal nature and loading (Fig. 6). The conversions oscillated between 68% and 87% in SBA-15-based materials and the range of conversion was smaller if MCF-based catalysts were applied (73–83%). However, the order of conversion values was irrespective of the support structure. Cerium(IV) oxide in the support caused a decrease in activity for both cerium loadings. In contrast, the effect of modification with niobium depended on its loading. For the lower niobium loading (Si/Nb = 103), a decrease in the activity was observed, whereas for the higher loading (Si/Nb = 19), an increase in benzaldehyde conversion took place with respect to the activity of materials without metals. The presented features could not be caused by the amount of imidazole (active phase) incorporated, because it was lower for e.g. Nb/SBA-15-19-Im than for Nb/SBA-15-103-Im (Table 1), whereas the activity of the first catalyst was the highest. Therefore, one can postulate the participation of niobium species in the reaction pathway, similarly as it has been indicated in ref. 14. Niobium containing materials exhibited Lewis acidity, as indicated by pyridine adsorption (Fig. 3). Interestingly, in Nb/MCF-19, the amount of LAS was ca. 1.7 times higher than in Nb/MCF-103 after evacuation at 150 °C and 200 °C. Lewis acidity is important for the Knoevenagel condensation. It is known that acid–base cooperation plays a positive role in promoting the Knoevenagel condensation via the ion-pair mechanism.43–47 Basic sites abstract a proton from the methylene carbon forming a carbanion. The carbonyl group of the aldehyde is activated by acidic sites and a carbocation is generated. The supports modified by CeO2 did not have Lewis acid sites and therefore did not enhance the catalyst activity.


image file: c7cy00927e-f6.tif
Fig. 6 The role of metal (Nb or Ce) in the support on the conversion of benzaldehyde in the Knoevenagel condensation performed at different pKa. Conversion of benzaldehyde after 240 min for pKa = 9; 240 min for pKa = 10.7; and 300 min for pKa = 13.3.

It is important to add that the selectivity of the reaction between benzaldehyde and ethyl cyanoacetate was 100% to the main product, ethyl-(2E)-2-cyano-3-phenyl-2-propenoate. It was not the case for the reaction performed between benzaldehyde and ethyl acetoacetate (pKa = 10.7). In the latter reaction carried out over SBA-15-based catalysts, the metals in the support only negligibly influenced the benzaldehyde conversion. In contrast, for the MCF structure, the activity was strongly dependent on metal loading; the higher the loading, the lower the activity, irrespective of the metal nature. It is worth noting that in this reaction, cerium-containing catalysts were more active than niobium modified materials. It suggests that the redox/basic properties of CeO2 were more important for this reaction than Lewis acidity coming from niobium species. In fact, the activity of the supports (examples for MCF) containing metals without imidazole was higher if cerium was used as modifier instead of niobium species (Table S3-SD). The presence of imidazole enhanced the activity but especially significantly increased the selectivity to the main product.

Looking for the reasons for the above described features, the selectivity of these reactions was analyzed in detail. Fig. 7 displays the curves of benzaldehyde conversion as well as selectivity to the main product (2-acetyl-3-phenylacrylic acid ethyl ester) and other products obtained over the catalysts with the lower metal loading. The conversion of benzaldehyde systematically increased with the reaction time. At the beginning of the reaction, at low benzaldehyde conversion, the selectivity to other products was relatively high, although the main product dominated. With the reaction time, the selectivity to the main product increased, whereas that to the other products decreased.


image file: c7cy00927e-f7.tif
Fig. 7 Activity (■) and selectivity (●- main product and ▲- others products) in the Knoevenagel condensation between benzaldehyde and ethyl acetoacetate (pKa = 10.7).

Interestingly, in MCF materials modified with metals after ca. 60 min of the reaction, the selectivity to the other products increased again and that to the main product decreased. This effect was especially pronounced for the cerium-containing material. Such a behavior indicated changes in the reaction pathways during the reaction, which influenced selectivity but not activity. Comparison of selectivities for almost the same benzaldehyde conversion is shown in Fig. 8. The highest selectivity to the main product was achieved over the MCF structure with the high metal loading (Si/metal = 19), irrespective of the nature of the metal. The presence of Nb or Ce species played an important role in both activity and selectivity of the reaction. It was confirmed by the reaction curves plotted for the process performed over SBA-15-Im and MCF-Im, i.e. the catalyst without any metal (Fig. 7). At low benzaldehyde conversion, the selectivity to the other products significantly dominated over the selectivity to the main product on the catalyst without any metal, in contrast to the catalysts based on the supports containing metals. With the reaction time, the selectivity to the main product increased, whereas that to the other products decreased. However, for the whole reaction time, the selectivity to the main product was much lower over SBA-15-Im and MCF-Im than over the catalysts containing metals in the support (Nb/SBA-15-103-Im, Ce/SBA-15-103-Im, Nb/MCF-103-Im and Ce/MCF-103-Im).


image file: c7cy00927e-f8.tif
Fig. 8 Selectivity in the Knoevenagel condensation between benzaldehyde and ethyl acetoacetate (pKa = 10.7) – comparison for almost the same benzaldehyde conversion obtained on imidazole-modified catalysts.

In the reaction with diethyl malonate (pKa = 13.3), the highest activity was achieved over the catalysts containing cerium (high loading: Si/Ce = 19), irrespective of the support structure (Fig. 6). The yield/selectivity of products depended on the catalyst structure (Table S4 and Fig. 9). On cerium-containing (Si/Ce = 103) catalysts, the yield of the main product was lower than the yield of other products if the SBA-15 structure was used, in contrast to the MCF structure. The other products can be derived from secondary reactions such as Michael addition of benzaldehyde and diethyl malonate, oxidation of benzaldehyde to benzoic acid and/or diethyl malonate auto-condensation (Scheme S1-SD). If only the structure (shape selectivity effect) was responsible for these features, the same phenomenon should be noted for niobium-containing catalysts. However, the mentioned behavior was not the case for the catalysts with niobium. Therefore, one has to consider not only the effect of the structure but also the role of cerium(IV) oxide in the formation of side reaction products. CeO2 is a well-known active species in oxidation processes.48,49 Thus, it can activate oxidation of benzaldehyde to benzoic acid which is much easier formed than the main product due to the shape selectivity effect. The highest yield of the main product was reached using Ce/MCF-Im-19. The basic mobile oxygen in CeO2 has to take part in this reaction besides the imidazole active species. In contrast, acidic niobium species, which also promoted condensation between benzaldehyde and diethyl malonate, exhibited a poorer promoting effect than cerium species.


image file: c7cy00927e-f9.tif
Fig. 9 Yields of the main product (■) and other products (●) in the reaction between benzaldehyde and diethyl malonate (pKa = 13.3) performed on the catalysts modified with imidazole and based on metal-containing supports.

The stability of all catalysts in all Knoevenagel condensations was estimated by analysis of the active phase leaching. The selected solutions after the reactions were subjected to GC-MS analyses. No solution contained the active basic species used as modifiers of the catalysts. This means that leaching of the active phases into the reagent mixtures did not occur.

3.2.3. The role of the type of nitrogen-containing organic modifier. The effect of the modifier type (imidazole, triazole and aminotriazole – Scheme 1) anchored on MCF and NbMCF on the activity in Knoevenagel condensation was estimated on the basis of the reaction between benzaldehyde and ethyl cyanoacetate (pKa = 9). As the amounts of modifiers were different, depending on their type and the chemical composition of the support (MCF, NbMCF) (Table 1), the comparison of activity was made on the basis of the TOF (turnover frequency) calculated for each reaction. Fig. 10 displays the results obtained. It is clear that, irrespective of the support content, triazole anchored to the support was much less active than the imidazole species. The three nitrogen atoms in the triazole ring were responsible for its lower basicity than the two nitrogen atoms in the imidazole ring. Interestingly, if an amine group was anchored on the triazole ring via a click reaction to form 1,2,3-triazol-4-ylmethanamine, the modifier formed presented higher activity than that of triazole species but lower than that of imidazole. The amine group enhanced benzaldehyde conversion. Niobium incorporated to the silica structure in the MCF material decreased the TOF values, irrespective of the type of nitrogen modifier. The reason is that niobium species (LAS) interacted with the basic modifiers and in this way lowered the basicity of nitrogen in the modifiers.
image file: c7cy00927e-f10.tif
Fig. 10 Activity of catalysts containing different nitrogen modifiers in the condensation between benzaldehyde and ethyl cyanoacetate (pKa = 9), expressed as turnover frequency (TOF) calculated as mole of benzaldehyde converted on one mole of nitrogen-containing compounds (imidazole, triazole, or triazolamine).

4. Discussion

Two different metal modifiers (niobium and cerium species) of the ordered mesoporous silica supports were used as potential promoters in the Knoevenagel condensation between benzaldehyde and methylenic compounds carried out over the catalysts containing imidazole, or triazole, or aminotriazole (1,2,3-triazol-4-ylmethanamine). The fundamental difference between the two metal modifiers of the support was their location in the mesoporous structure. Niobium was included into the silica matrices, whereas cerium was placed as crystalline CeO2 in pores and cells of the SBA-15 and MCF materials. The location and chemical properties of both types of modifiers influenced the surface properties of the supports. Niobium-containing supports exhibited Lewis acidity, which was a consequence of Nb inclusion into the silica skeleton, as discussed in ref. 36, whereas cerium modified supports showed properties typical of CeO2. In most catalytic reactions, CeO2 plays the role of an oxygen buffer by storing and releasing O2 due to the Ce4+/Ce3+ redox couple, which results from the intrinsic defect chemistry of ceria (Ce3+ and oxygen vacancy concentration).49 Thus, ceria was used in several oxidation reactions as the active phase or the support for metals. However, it was also widely applied in C–C coupling reactions like aldol condensation or Michael condensation.49 In these reactions, the ceria basicity is important; its basicity is lower than that of Ca(OH)2 but higher than that of ZrO2.50 Thus, two different properties were generated in the supports modified by metals. The niobium-containing supports developed Lewis acidity, whereas the supports containing CeO2 exhibited basic/redox properties.

The role of metallic modifiers in the supports can be considered on the basis of the three Knoevenagel condensations (pKa = 9, pKa = 10.7 and pKa = 13.3) performed on the catalysts containing imidazole as the active species (Fig. 6). With decreasing pKa of the methylene compound (from 13.3 to 9), the role of acidity of the support coming from niobium species increased. For the condensation between benzaldehyde and ethyl cyanoacetate (pKa = 9), Nb/MCF-Im-19 and Nb/SBA-15-Im-19 were the most active catalysts. The reaction gave rise only to the main product (ethyl-(2E)-2-cyano-3-phenyl-2-propenoate). The participation of Lewis acid sites in the Knoevenagel condensations was discussed in earlier papers.14,45 In contrast to the niobium-containing catalysts, the role of CeO2 in the enhancement of activity increased with increasing pKa (from 9 to 13.3) of the methylene compound taking part in the Knoevenagel condensation. The cerium-containing catalysts led to higher conversion of benzaldehyde in the reaction with ethyl acetoacetate (pKa = 10.7), but the role of CeO2 was especially significant in the condensation between benzaldehyde and diethyl malonate (pKa = 13.3). The highest activity in this reaction was observed for the catalysts with the higher loading of cerium species (Si/Ce = 19), irrespective of the structure. Thus, the role of ceria as a promoter that enhances the catalysts activity was evident. Interestingly, the yield of the side reaction products was significant and is dependent on the catalyst structure. In the reaction performed on Ce/SBA-15-103-Im, the yield of the main product was lower than that of the other products. If the MCF structure was used, the opposite yield relation was observed i.e. the domination of the main product. Therefore, one can conclude that ceria in the catalysts activated the reaction of benzaldehyde oxidation rather than the side reactions, because the bulky product of the latter reaction (Scheme S1-SD) should be easily formed in MCF materials, whereas the side reaction products dominated over SBA-15. The effect of the structure was also noted for the reaction at pKa = 10.7. The structure impacted both activity and selectivity of the reaction in contrast to the reaction at pKa = 13.3 in which only selectivity was structure dependent. In the reaction between benzaldehyde and ethyl acetoacetate, many different side reaction products can be formed (Scheme S2-SD). Such products were formed at the beginning of the reaction at low benzaldehyde conversion. With increasing conversion, the reaction was directed to the main product. However, in MCF materials after some time of the reaction, the selectivity to the main product decreased and that to the side reaction products increased. This effect was more pronounced on Ce/MCF-Im because ceria activated aldol condensation49 towards the bulky product, which can easy diffuse only in large pores of cellular foams (MCF). In the smaller pores of SBA-15, this reaction pathway seemed to be blocked.

The results concerning the role of nitrogen compounds used as basic modifiers indicated that the relationship between the number of nitrogen atoms in the basic modifier and the activity of the catalysts supported on the MCF structure and the activity in the reaction between benzaldehyde and ethyl acetoacetate was not simple. Not only the number of nitrogen atoms but also the structure and nature of nitrogen species (imine, amine, etc.) played important roles. The amine group in 1,2,3-triazol-4-ylmethanamine loaded on MCF was expected to exhibit the highest activity in Knoevenagel condensation. In the interim, the use of imidazole species led to higher conversion of benzaldehyde. Thus, the interaction between the amine group and a nitrogen in the triazole ring, discussed in the previous sections, caused a decrease in the basicity of nitrogen species.

Finally, it is worth noting that niobium present in the supports enhanced the thermal stability of the nitrogen compounds anchored on the supports' surface by the chemical interaction between LAS (Nb species) and basic sites (nitrogen compound). Cerium did not show such a behavior because redox/basic CeO2 did not interact chemically with the basic modifier (nitrogen compounds).

5. Conclusions

Imidazole active centers located in mesoporous SBA-15 and MCF materials exhibited the highest efficiency, among the active nitrogen compounds used in this work, in all Knoevenagel condensations tested. The increase in the number of nitrogen atoms in the active phases weakened their basicity, which resulted in the lower activity in the reaction between benzaldehyde and ethyl cyanoacetate (pKa = 9).

The modifiers included into the supports influenced the activity and selectivity in Knoevenagel condensations performed with methylenic compounds at different pKa. CeO2 present in the supports caused an increase in the activity of imidazole catalysts and selectivity to the side reaction products with increasing pKa of the methylene compounds. The opposite relation was observed for the niobium-containing catalysts. Niobium Lewis acid sites took part in the main reaction pathway increasing the benzaldehyde conversion. This effect increased with lowering of pKa of the methylene compound used in Knoevenagel condensation. Thus, the cerium-containing imidazole catalysts exhibited the highest activity in the reaction between benzaldehyde and diethyl malonate (pKa = 13.3), whereas the niobium containing catalysts presented the highest activity in the condensation with ethyl cyanoacetate (pKa = 9).

The catalyst's structure influenced the activity and selectivity in the reaction with diethyl malonate and only the activity in the reaction with ethyl acetoacetate. In the reaction at pKa = 13.3 performed at 150 °C on the SBA-15-based imidazole catalyst, the oxidation of benzaldehyde was faster than the main Knoevenagel condensation reaction because of the steric effect limiting the rate of the main reaction route. In contrast, the MCF structure allowed the main reaction pathway to easily proceed. The structure effect for the reaction at pKa = 10.7 was due to the promotion of the condensation reaction in bigger pores of MCF materials.

Acknowledgements

The National Science Centre of Poland (Grant No. 2014/15/B/ST5/00167) is acknowledged for its financial support. This work has been also supported by the Spanish Program of Research, Development and Innovation [Project CTM2014-56668-R]. VCC acknowledges UNED for her postdoctoral contract.

References

  1. J. Mondal, A. Modak and A. Bhaumik, J. Mol. Catal. A: Chem., 2011, 335, 236–241 CrossRef CAS .
  2. Sh.-W. Bain, Z. Ma, Z.-M. Cui, L.-Sh. Zhang, F. Niu and W.-G. Song, J. Phys. Chem. C, 2008, 112, 11340–11344 CAS .
  3. J. Puriwat, W. Chaitree, K. Suriye, S. Dokjampa, P. Praserthdam and J. Panpranot, Catal. Commun., 2010, 12, 80–85 CrossRef CAS .
  4. X. Wang, K. S. K. Lin, J. C. C. Chan and S. Cheng, J. Phys. Chem. B, 2005, 109, 1763–1769 CrossRef CAS PubMed .
  5. A. E. Palomares, G. Eder-Mirth, M. Rep and J. A. Lercher, J. Catal., 1998, 180, 56–65 CrossRef CAS .
  6. K. Wilson and A. F. Lee, Catal. Sci. Technol., 2012, 2, 884–897 CAS .
  7. J. Marco-Contelles, E. Pérez-Mayoral, A. Samadi, M. C. Carreiras and E. Soriano, Chem. Rev., 2009, 109, 2652–2671 CrossRef CAS PubMed .
  8. M. Godino-Ojer, E. Soriano, V. Calvino-Casilda, F. J. Maldonado-Hódar and E. Pérez-Mayoral, Chem. Eng. J., 2013, 314, 488–497 CrossRef .
  9. M. Polozij, E. Pérez-Mayoral, J. Cejka, J. Hermann and P. Nachtigall, Catal. Today, 2013, 204, 101–107 CrossRef CAS .
  10. L.-B. Sun, X.-Q. Liu and H.-C. Zhou, Chem. Soc. Rev., 2015, 44, 5092–5147 RSC .
  11. Z. Zienkiewicz, V. Calvino-Casilda, I. Sobczak, M. Ziolek, R. M. Martin-Aranda and A. J. Lopez-Peinado, Catal. Today, 2009, 142, 303–307 CrossRef CAS .
  12. J. H. Oh, B. S. Ahn, J.-S. Han, S. D. Lee, S. W. Kim and H. Lee, Bull. Korean Chem. Soc., 2008, 29, 2202–2204 CrossRef CAS .
  13. T. M. Suzuki, T. Nakamura, K. Fukumoto, M. Yamamoto, Y. Akimoto and K. Yano, J. Mol. Catal. A: Chem., 2008, 280, 224–232 CrossRef CAS .
  14. V. Calvino-Casilda, M. Olejniczak, R. M. Martin-Aranda and M. Ziolek, Microporous Mesoporous Mater., 2016, 224, 201–207 CrossRef CAS .
  15. J. M. Maestre, J. P. Sarasa, C. Bo and J. M. Poblet, Inorg. Chem., 1998, 37, 3071–3077 CrossRef CAS .
  16. D. Blasco-Jiménez, I. Sobczak, M. Ziolek, A. J. López-Peinado and R. M. Martín-Aranda, Catal. Today, 2010, 152, 119–125 CrossRef .
  17. X.-Y. Liu, L.-B. Sun, F. Lu, T.-T. Li and X.-Q. Liu, J. Mater. Chem. A, 2013, 1, 1623–1631 CAS .
  18. X.-Y. Liu, L.-B. Sun, X.-D. Liu, A.-G. Li, F. Lu and X.-Q. Liu, ACS Appl. Mater. Interfaces, 2013, 5, 9823–9829 CAS .
  19. L. Zhu, F. Lu, X.-D. Liu, X.-Q. Liu and L.-B. Sun, Chem. Commun., 2015, 51, 10058–10061 RSC .
  20. L.-B. Sun, X.-Y. Liu, A.-G. Li, X.-D. Liu and X.-Q. Liu, Chem. Commun., 2014, 50, 11192–11195 RSC .
  21. N. Lingaiah, P. S. S. Prasad, P. K. Rao, L. E. Smart and F. J. Berry, Appl. Catal., A, 2001, 213, 189–196 CrossRef CAS .
  22. X. Zhang, E. S. Man Lai, R. Martin-Aranda and K. L. Yeung, Appl. Catal., A, 2004, 261, 109–118 CrossRef CAS .
  23. N. Aider, A. Smuszkiewicz, E. Pérez-Mayoral, E. Soriano, R. M. Martín-Aranda, D. Halliche and S. Menad, Catal. Today, 2014, 227, 215–222 CrossRef CAS .
  24. D. Kryszak, K. Stawicka, V. Calvino-Casilda, R. Martin-Aranda and M. Ziolek, Appl. Catal., A, 2017, 531, 139–150 CrossRef CAS .
  25. K. Stawicka, K. Drazkiewicz and M. Ziolek, Microporous Mesoporous Mater. Search PubMed  submitted.
  26. B. Kilos, A. Tuel, M. Ziolek and J.-C. Volta, Catal. Today, 2006, 118, 416–424 CrossRef CAS .
  27. M. Trejda, A. Tuel, J. Kujawa, B. Kilos and M. Ziolek, Microporous Mesoporous Mater., 2008, 110, 271–278 CrossRef CAS .
  28. D. C. Tranca, A. Wojtaszek-Gurdak, M. Ziolek and F. Tielens, Phys. Chem. Chem. Phys., 2015, 17, 22401–22411 RSC .
  29. P. Kaminski and M. Ziolek, Appl. Catal., B, 2016, 187, 328–341 CrossRef CAS .
  30. P. Kaminski, M. Ziolek, B. Campo and M. Daturi, Catal. Today, 2015, 243, 218–227 CrossRef CAS .
  31. L. Escamilla-Perea, R. Nava, B. Pawelec, M. G. Rosmaninho, C. L. Peza-Ledesma and J. L. G. Fierro, Appl. Catal., A, 2010, 381, 42–53 CrossRef CAS .
  32. B. Chakraborty and B. Viswanathan, Catal. Today, 1999, 49, 253–260 CrossRef CAS .
  33. C. A. Emeis, J. Catal., 1993, 141, 347–354 CrossRef CAS .
  34. Y.-M. Liu, J. Xu, L.-C. Wang, Y. Cao, H.-Y. He and K.-N. Fan, Catal. Lett., 2008, 125, 62–68 CrossRef CAS .
  35. S. Khabtou, T. Chevreau and J. C. Lavalley, Microporous Mater., 1994, 3, 133–148 CrossRef CAS .
  36. M. Ziolek and I. Sobczak, Catal. Today, 2017, 285, 211–225 CrossRef CAS .
  37. M. Thommes, K. Kaneko, A. V. Neimark, J. P. Olivier, F. Rodriguez-Reinoso, J. Rouquerol and K. S. W. Sing, Pure Appl. Chem., 2015, 87, 1051–1069 CrossRef CAS .
  38. M. Olejniczak and M. Ziolek, Microporous Mesoporous Mater., 2014, 196, 243–253 CrossRef CAS .
  39. R.-Sh. Juang, Ch.-H. Liang, W.-Ch. Ma, Ch.-Y. Tsai and Ch. Huang, J. Taiwan Inst. Chem. Eng., 2014, 45, 3046–3051 CrossRef CAS .
  40. B. Stuart, Infrared Spectroscopy: Fundamentals and Applications, John Wiley & Sons, 2004 Search PubMed .
  41. S. A. Lawrence, Amines, Synthesis, Properties and Applications, Cambridge University Press, Cambridge, 2004 Search PubMed .
  42. A. Wach, M. Drozdek, B. Dudek, P. Łątka and P. Kuśtrowski, Microporous Mesoporous Mater., 2016, 226, 433–443 CrossRef CAS .
  43. M. J. Climent, A. Corma, S. Iborra and A. Velty, J. Mol. Catal. A: Chem., 2002, 327, 182–183 Search PubMed .
  44. Y. Q. Shao, J. Q. Guan, S. J. Wu, H. Liu, B. Liu and Q. B. Kan, Microporous Mesoporous Mater., 2010, 128, 120–127 CrossRef CAS .
  45. F. Vermoortele, R. Ameloot, A. Vimont, C. Serre and D. D. Vos, Chem. Commun., 2011, 47, 1521–1523 RSC .
  46. S. L. Hruby and B. H. Shanks, J. Catal., 2009, 263, 181–188 CrossRef CAS .
  47. A. Corma, V. Fornés, R. M. Martín-Aranda, H. García and J. Primo, Appl. Catal., 1990, 59, 237–248 CrossRef CAS .
  48. P. Kaminski, M. Ziolek and J. A. van Bokhoven, RSC Adv., 2017, 7, 7801–7819 RSC .
  49. A. Trovarelli and P. Fornasiero, Catalysis by ceria and related materials, Imperial College Press, London, 2013 Search PubMed .
  50. S. Lippert, W. Baumann and K. Thomke, J. Mol. Catal., 1991, 69, 199–214 CrossRef CAS .

Footnote

Electronic supplementary information (ESI) available: Procedures of catalysts' syntheses. Methods of catalyst characterization. Reaction conditions in Knoevenagel condensations. Schemes of the Knoevenagel condensations. Texture/structure parameters of catalysts. Conversion of benzaldehyde and selectivities in Knoevenagel condensations. UV-vis spectra and DTA-TG results. See DOI: 10.1039/c7cy00927e

This journal is © The Royal Society of Chemistry 2017