Abdelnasser Abidli*ab
aDepartment of Soil Sciences and Agri-Food Engineering, Laval University, G1V 0A6, Quebec City, Quebec, Canada. E-mail: abdelnasser.abidli.1@ulaval.ca; Fax: +1-418-656-3723; Tel: +1-581-777-0860
bCentre in Green Chemistry and Catalysis (CGCC), H3A 0B8, Montreal, Quebec, Canada
First published on 22nd October 2015
Hexagonally well-organized ZnCl2-OMA materials have been successfully synthesized by a one-pot approach using a sol–gel method at ambient temperature and pressure. These materials were prepared through simultaneous self-assembly process with F127 as directing agent combined with in situ impregnation of ZnCl2. The new approach was found to be compatible with various common aluminium precursors and carboxylic acids used as self-assembly interfacial protectors. The obtained ZnCl2-OMA materials were fully characterized using XRD, N2 adsorption–desorption, TEM, SEM, EDX, XPS and 1H and 27Al MAS-NMR techniques. These materials were compared with the zinc chloride-doped organized mesoporous alumina (ZnCl2-OMA) supports prepared through a conventional two steps process that includes OMA synthesis and then post-synthesis incorporation of ZnCl2. Thus, the one-pot synthesized materials were found to exhibit improved properties compared to the conventional ones, particularly larger BET surface area. The synthesized ZnCl2-OMA materials were then used as catalytic supports for methyltrioxorhenium (MTO), showing enhanced catalytic performance for methyl oleate self-metathesis, demonstrating better activity and selectivity towards desired metathesis products. These features are highly beneficial for large-scale materials synthesis through fast, simple, easy and low-cost introduction of functionalities on mesoporous materials surface without multi-step procedures.
Therefore, the most suitable route for the modification and functionalization of these catalyst supports is to perform an in situ incorporation of the desired components, prior (or during) to the formation of the host materials. These one-step procedures will highly increase the efficiency of the synthesis processes by offering a variety of advantages including faster synthesis by shortening the process from several days to few hours or minutes.18–22 These one-step approaches are more cost-effective and leads to higher reaction yields.23 This renders the introduction of desired functionalities easier and simpler with limited pore blocking phenomena.24 Moreover, this methodology offers better control over the materials synthesis, uniformity, purity and homogeneity.22,25,26 Thus, allowing access to catalytic supports with better properties that are key factor for the enhancement of their catalytic performance.18,27 Actually, several reports showed that such synthesized catalysts always exhibit significantly better catalytic properties compared to those prepared via the conventional multistep process.18,27,28
Therefore, in view of these promising features, we decided to adopt this methodology for our study on modified mesoporous alumina-based metathesis catalysts. Previously, we have reported a successful synthesis of a highly active metathesis catalysts using methyltrioxorhenium (MTO) supported on functionalized mesoporous alumina. First, through several optimizations we showed that ZnCl2 is the most suitable metal halide for alumina modification, offering the highest metathesis performance among several other metal halides used for the study.29–31 The MTO supported ZnCl2-modified Al2O3 catalyst was found to exhibit higher metathesis activity compared to the unmodified alumina.29–31 This catalyst showed high performance for both methyl oleate and triolein self-metathesis.29–31 However, this catalyst displayed relatively slow kinetics and unsatisfactory selectivity towards desired metathesis products. Therefore, we replaced the traditional alumina support having wormhole-like mesostructure by a hexagonally ordered mesoporous alumina (OMA). Thus, we have recently reported an enhanced performance of the MTO-based catalysts for methyl oleate self-metathesis using the ZnCl2-modified OMA supports.32 The OMA-based catalysts were successfully tested for their activity for methyl oleate self-metathesis. In fact, metathesis reaction in one of the most interesting catalytic route providing access to a variety of valuable monomers via CC bond formation.33,34 On the other hand, methyl oleate is chosen as a model molecule having both long carbon chain containing C
C bond which will be subjected to metathesis catalysis conditions, and also the ester functional groups to evaluate the MTO-based catalyst tolerance for functional groups. In addition, methyl oleate is a bulky molecule which is suitable to evaluate the kinetics and molecular diffusion inside the mesoporous network. In the meantime, this renewable substrate originated form vegetable oils represents an interesting feedstock for the synthesis of value-added compounds via valorization through metathesis pathway.35,36
Therefore, in this study, in order to improve furthermore the MTO-based catalyst design and performance, we developed a new direct one-pot self-assembly strategy for the synthesis of ZnCl2-modified OMA supports. Instead of the traditional methodology that we usually adopted which consist on preparing the OMA supports then a post-synthetic functionalization with ZnCl2 in a laborious process. Thus, we are aiming to improve the synthesis process trough a simpler, faster, easier and cost-effective process. The new synthesis protocol is expected to offer the desired materials with enhanced chemical, morphological, and textural properties which may likely lead to better metathesis catalytic performance. In addition, higher synthesis yields are expected to be reached. Detailed characterizations and comparison with alumina-based supports prepared via the conventional approach and their catalytic performance are provided. This successful methodology is a step forward towards an easy and efficient scalability. Also provide a rapid access route for such desired materials, especially with the recent attention gathered around zinc chloride impregnated on alumina and alumina–silica materials for other applications such as methyl chloride synthesis37,38 and catalytic alkylation reactions.39–42 As well as other zinc chloride-modified materials including mesoporous carbon nitride (ZnCl2/mp-C3N4) and doped graphitic carbon nitride (Zn-g-C3N4) for other catalytic applications (e.g. cycloaddition and transesterification).43,44 Therefore, in addition to the synthesis of modified-alumina, this method represents a promising approach to prepare various functionalized mesoporous materials for broad applications.
![]() | ||
Fig. 1 (A) Nitrogen adsorption–desorption isotherms and (B) pore size distributions of the prepared ZnCl2-OMA samples. All samples were calcined at 400 °C. |
Table 1 summarizes the Brunauer–Emmet–Teller (BET) surface areas (SBET), total pore volumes (PV) average pore size (dBJH) data obtained by N2 adsorption–desorption at 77 K illustrating the textural properties and porosity parameters of the synthesized ZnCl2-OMA materials. Relatively large BET surface are and pore volume were obtained, up to 358 m2 g−1 and 0.67 cm3 g−1, respectively. In addition, nearly all the samples displayed uniform narrow pore size distribution with a mean diameter from 3.4 to 6.5 nm. However, some analyzed samples present some secondary porosity (Table 1). These results suggest that the presence of ZnCl2 did not hamper the formation of the desired hexagonally ordered mesoporous materials.
ZnCl2-OMA samplesa | Acid | Aluminium precursor | Synthesis approach | SBETb (m2 g−1) | PVc (cm3 g−1) | dBJHd (nm) |
---|---|---|---|---|---|---|
a Al![]() ![]() ![]() ![]() |
||||||
ZnCl2-OMA-1 | Tartaric | Al(OBus)3 | One-pot | 203 | 0.30 | 3.8 and 5.6 |
Two-stepe | 154 | 0.26 | 3.8 | |||
ZnCl2-OMA-2 | Fumaric | Al(OBus)3 | One-pot | 299 | 0.41 | 3.8 |
Two-stepe | 214 | 0.31 | 3.8 | |||
ZnCl2-OMA-3 | Citric | Al(OBus)3 | One-pot | 307 | 0.42 | 3.8 |
Two-stepe | 316 | 0.44 | 3.8 | |||
ZnCl2-OMA-4 | Oxalic | Al(OBus)3 | One-pot | 282 | 0.36 | 3.8 |
Two-stepe | ND | ND | ND | |||
ZnCl2-OMA-5 | Maleic | Al(OBus)3 | One-pot | 238 | 0.27 | 3.4 |
Two-stepe | ND | ND | ND | |||
ZnCl2-OMA-6 | Malonic | Al(OBus)3 | One-pot | 252 | 0.27 | 3.4 |
Two-stepe | ND | ND | ND | |||
ZnCl2-OMA-7 | Oxalic | Al(OPri)3 | One-pot | 234 | 0.46 | 3.8 and 6.5 |
Two-stepe | ND | ND | ND | |||
ZnCl2-OMA-8 | Malonic | Al(OPri)3 | One-pot | 181 | 0.23 | 3.4 |
Two-stepe | ND | ND | ND | |||
ZnCl2-OMA-9 | Acetic | Al(OPri)3 | One-pot | 199 | 0.32 | 3.8 and 4.9 |
Two-stepe | ND | ND | ND | |||
ZnCl2-OMA-10 | Maleic | Al(OPri)3 | One-pot | 159 | 0.34 | 3.8 and 6.5 |
Two-stepe | ND | ND | ND | |||
ZnCl2-OMA-11 | Tartaric | Al(OPri)3 | One-pot | 219 | 0.38 | 3.8 and 4.9 |
Two-stepe | 190 | 0.39 | 5.7 | |||
ZnCl2-OMA-12 | Citric | Al(OPri)3 | One-pot | 232 | 0.47 | 7.7 |
Two-stepe | 241 | 0.38 | 3.8 and 4.9 | |||
ZnCl2-OMA-13 | Citric | Al(OBut)3 | One-pot | 358 | 0.67 | 3.8 and 5.6 |
Two-stepe | 239 | 0.41 | 4.9 | |||
ZnCl2-OMA-14 | Tartaric | Al(OBut)3 | One-pot | 307 | 0.44 | 3.8 |
Two-stepe | 257 | 0.44 | 5.6 | |||
ZnCl2-OMA-15 | Malonic | Al(OBut)3 | One-pot | 224 | 0.41 | 3.8 |
Two-stepe | ND | ND | ND | |||
ZnCl2-OMA-16 | Citric | Al(NO3)3·9H2O | One-pot | 201 | 0.39 | 3.8 and 4.9 |
Two-stepe | 139 | 0.40 | 3.8 and 6.5 | |||
ZnCl2-OMA-17 | Tartaric | Al(NO3)3·9H2O | One-pot | 226 | 0.41 | 3.8 and 5.6 |
Two-stepe | 222 | 0.46 | 3.8 | |||
ZnCl2-OMA-18 | Malonic | Al(NO3)3·9H2O | One-pot | 183 | 0.33 | 3.8 |
Two-stepe | ND | ND | ND |
As mentioned above, the one-pot synthesis was performed using distinct aluminium precursors and carboxylic acids in order to generalize the synthesis procedure and verify it's reproducibly.
Therefore, we used the same precursors (Al(OBus)3, Al(OPri)3, Al(OBut)3 and Al(NO3)3·9H2O) and acids (citric, tartaric, fumaric, oxalic, malonic, maleic and acetic acid) that we have previously used for the synthesis of catalytic ZnCl2-OMA materials through conventional two-steps process.32
As depicted in Fig. 1 and Table 1, we successfully prepared various ZnCl2-OMA samples using different carboxylic acids, while no significant difference was observed on their textural properties and their porosity parameters. Furthermore, the targeted organized ZnCl2-OMA materials were successfully obtained upon the use of different aluminium precursors as illustrated in Table 1 and Fig. 2 with no significant shift. This similarity is ascribed to the identical behavior of the investigated precursors and acids, suggesting that the different carboxylic acids tested exhibit similar coordination abilities with the metal centers (Al) via monodentate or bridging bidentate modes during the mesophase formation pathway.32,47 On the other hand, the investigated aluminium precursors proved also similar chelation abilities of their metallic centers (Al) with both the templating agent (F127) fragments and the added carboxylic acids used as organic–inorganic interfacial protectors assisting the ordered self-assembly process. This kind of organized assembly is tailored by means of several interfacial interactions including covalent and hydrogen bonding as well as van der Waals interactions, which plays a pivotal role in the mesophase construction.48,49
![]() | ||
Fig. 2 (A) Nitrogen adsorption–desorption isotherms and (B) BJH pore size distributions curves of the prepared ZnCl2-OMA sample. All samples were calcined at 400 °C. |
In comparison with the materials synthesized via the conventional two-steps process,32 as depicted in Table 1, almost all the newly prepared ZnCl2-OMA samples (one-pot) possess larger BET surface area. As we reported previously, we believe that the lower BET surface area obtained through the two-steps synthesis is primarily ascribed to the additional calcination step during the ZnCl2 impregnation, as well as the filling of added ZnCl2 inside the already formed mesopores.32 These troublesome effects were advantageously avoided when adopting the one-pot synthesis process were in situ impregnation of ZnCl2 is occurred with no need for subsequent annealing step that may alter the materials surface.
Meanwhile, both synthesis procedures led to ZnCl2-OMA materials with relatively similar total pore volume and pore size distributions (Table 1), suggesting that independently from the synthesis process, similar porosity features can be obtained using the same aluminium precursors and carboxylic acids having identical ligands and carbon chain length, respectively.
![]() | ||
Fig. 3 Small-angle X-ray scattering (SAXS) patterns of zinc chloride-modified ordered mesoporous alumina samples. All samples were calcined at 400 °C. |
As illustrated from the BET results, similarly, the XRD analysis revealed no significant shift in the XRD patterns when using distinct carboxylic acids (Fig. 3i) or different aluminium precursors (Fig. 3ii). It is worth to note that the obtained XRD analysis results were similar to those obtained for the ZnCl2-OMA samples prepared using the conventional two-steps process where the 2-D hexagonal p6mm structure was also detected for the prepared samples.32 Furthermore, no crystalline alumina phase was formed at 400 °C, as shown form the wide angle XRD patterns recorded for the calcined ZnCl2-OMA samples (Fig. S1 in the ESI†). These results indicate the formation of an amorphous network. More interestingly, the data obtained is a clear evidence for the absence of any additional Zn-containing crystalline phases (e.g. ZnAl2O4, ZnO), which are usually obtained in this range of calcination temperatures.52,53
In addition, the diameter of the cylindrical mesopores as observed from the TEM micrographs was around 5 nm in consistency with the obtained N2 adsorption–desorption data. The observed organized structure of the materials prepared through one-pot process was similar to that observed for the ZnCl2-OMA samples prepared using the conventional two-steps process.32 The TEM data confirm again that the presence of ZnCl2 during the mesophase formation does not alter the organized assembly of alumina during the templating and the calcination process, and preserves the organized alumina framework. Furthermore, the micrographs displayed in Fig. 4 indicate the absence of any aggregation of Zn species/phases.
To probe into the morphological features of the synthesized ZnCl2-OMA materials, the prepared samples were examined by SEM. Fig. 5a–d shows plane-view SEM images of the ZnCl2-OMA samples. The mesoporous ZnCl2-OMA microparticles exhibit a non-uniform three-dimensional (3D) architecture ranging from cuboid-like to plate-like shaped morphologies resulting from irregular micelles aggregation during the EISA process. Thus, the observed microparticles have a non-uniform particle size distribution varying from 20 to 50 μm with gaps between crystals (Fig. 5a–d). However, a relatively smooth particles surface is obtained, suggesting the excellent incorporation of ZnCl2. In our previous work, the ZnCl2-OMA materials prepared using the conventional two-steps process have smaller microparticles with less regular shape and morphology, which was ascribed to the mechano-chemical dislocation effect occurred during the ZnCl2 impregnation step.32 Unlike the two-steps conventional approach, the newly designed one-pot process led to efficient and homogeneous incorporation of ZnCl2 particle in the OMA matrix. However, we believe that the presence of ZnCl2 in the surfactant micelle before thermal treatment is among the reasons that led to relatively larger particles (compared to two-steps process) where ZnCl2 is entirely embedded in OMA phase.
Furthermore, as shown from previous analysis data (XRD, BET and TEM), the use of various carboxylic acids and aluminium precursors did not result in any significant morphology changes (Fig. 5a–d).
The Al 2p spectrum displayed in Fig. 6b revealed the presence of one component, the peak at 74.4 eV corresponds to Al present in the Al2O3 phase.54,55 Subsequently, the O 1s spectrum (Fig. 6c) can be deconvoluted into two components. Two peaks with similar intensities are observed at 531.2 eV and 532.5 eV, which are attributed to bulk oxygen present in the Al2O3 lattice (Al–O–Al) and higher concentration hydroxyl groups (Al–O–H) present in the Al2O3 materials surface, respectively.55–58 Unlike the conventionally prepared ZnCl2-OMA materials,32 the absence of a third peak at higher BE indicates that no molecular H2O is physisorbed at this temperature.58 These binding energy values are slightly higher compared to those we reported for the pure alumina materials.32 This could be due to the presence of ZnCl2 and its interactions with the species present in the alumina phase, which leads to a slight change in the electronic environment of these elements. The hydroxyl groups peak at 532.5 eV may also correspond to the Zn(OH)2 species that results from ZnCl2 interaction with alumina surface.59 The O 1s XPS data obtained previously for the two steps-prepared ZnCl2-OMA materials showed a lower intensity (amount) of Al–O–H peak compared to that of Al–O–Al.32 On the other hand, herein, the two peak intensities are quite equal. This could be due to elimination on the second calcination step which helps avoiding further dehydroxylation of alumina surface.
The zinc species bonding configurations were also investigated. The Zn 2p3/2 spectrum (Fig. 6d) shows one component with a sharp peak at 1022.1 eV, which could be assigned to distinct Zn2+ species including Zn–O, ZnCl2 or Zn(OH)2.32,60,61 We then studied the Cl− spectrum (Fig. 6e).
This spectrum can be deconvoluted into two components; a well-defined peak at 198.9 eV and a smaller peak at 200.5 eV correspond to Zn–Cl 2p3/2 and Zn–Cl 2p1/2 species, respectively.32,60,62 In addition to the possible Zn-containing species like bridging Al–O–Zn–O–Al or Zn(OH)2 species, the ZnCl2/Zn–Cl species are likely bonded to the alumina surface through interactions with surface hydroxyl groups. With the presence of Zn precursors in the reaction mixture before the formation of the alumina phase, we logically suspected the formation of other phases besides alumina. Zinc aluminate (ZnAl2O4) and zinc oxide (ZnO) are the most likely phases to be obtained under such similar conditions. Fortunately, the XPS measurements are clear evidence that no ZnAl2O4 phase was formed along with the ZnCl2-modified OMA phase. If ZnAl2O4 phase was present, the XPS analysis would show a single O 1s component associated with a peak at 531.4 eV, while Zn 2p3/2 and Zn 2p1/2 peaks would appear at 1044.9 eV and 1021.5–1021.9 eV, respectively.63,64 Similarly, no evidence of formation of zinc oxide phase (ZnO) was found, which is characterized by Zn 2p3/2 and Zn 2p1/2 peaks at 1021.5 eV and 1044.6 eV.64 It is worth to mention that even if zinc chloride thermal decomposition is possible, the resulted Zn is likely converted into the species cited above, while no trace of elemental Zn were found in the analyzed samples which can result in two Zn 2p3/2 and Zn 2p1/2 peaks at lower BE values; 1021.4 eV and 1044.5 eV, respectively, as reported previously.64
27Al MAS NMR experiments were conducted to investigate the Al environment and the changes related to the in situ ZnCl2 incorporation. Fig. 8 shows the various coordination geometries detected for Al species in the analyzed as-prepared (uncalcined) and calcined ZnCl2-OMA samples. A single sharp resonance peak was observed at around 0 ppm for the as-prepared sample (Fig. 8a). This suggests that the uncalcined sample mainly contains octahedral Al3+ species. This peak can be associated with the presence of the AlOOH boehmite intermediate phase,69,70 which will be then converted to alumina phase after annealing. Afterwards, the calcined ZnCl2-OMA samples were analyzed. The 27Al MAS NMR spectra displayed in Fig. 8b–e show a change of Al species coordination. We then observed the rise of one sharp and two broad resonance peaks (Fig. 8b–e). The major well-defined resonance peak centered at 3 ppm is attributed to the octahedral Al3+ species in the alumina phase (AlO6).70,71 Subsequently, the two broad minor resonance peaks centred at 32 and 63 ppm (Fig. 8b–e) correspond to the penta- and tetrahedral Al3+ species in the alumina phase (AlO5, AlO4), respectively.70 As reported recently in our work, this conversion of the octahedral Al3+ species upon calcination to pentahedral and tetrahedral Al3+ species is attributed to the partial substitution of oxygen ions by hydroxyl groups in octahedral Al centers during the calcination process.14,32,72 All 27Al MAS NMR spectra for the samples prepared using different carboxylic acids and diverse aluminium precursors show approximately the same profile regardless of the experimental conditions (Fig. 8, see also Fig. S6–S8 in the ESI†).
It is worth to mention that unlike alumina phase, the 27Al MAS NMR spectra of the ZnAl2O4 spinel obtained under similar thermal conditions (300–500 °C) exhibits a single peak around 0 ppm which is characteristic of Al3+ ions occupying AlO6 sites, as reported previously.52
In comparison, the ZnCl2-OMA samples prepared using the two step process resulted in AlO5 and AlO6 resonance peaks with relatively similar intensities.32 Herein, we observe that the AlO5 resonance peaks have lower intensities compared to those of the AlO6 resonance peaks (Fig. 8, see also Fig. S6–S8 in the ESI†), indicating the lower number of AlO5 sites. This suggests that the in the one-pot synthesized ZnCl2-OMA materials the zinc chloride was preferably doped on these sites. This suggestion is in a good agreement with the earlier speculation about the distribution of the Al3+ species in the alumina matrix; where it has been proposed that AlO6 sites are found in the bulk of the alumina inorganic framework walls, while the AlO4 and AlO5 sites are likely on the surface.71 A similar phenomenon was recently observed for the Cu–Al2O3 system.73
![]() | ||
Scheme 1 Schematic representation of the one-pot evaporation induced self-assembly process for the formation of ZnCl2 embedded in well-organized mesoporous alumina (OMA) materials. |
In order to highlight the advantages of the newly prepared ZnCl2-OMA materials as catalytic supports for MTO, here we present an extended comparison with other alumina-based catalytic support those we have previously synthesized and used for the design of the MTO metathesis catalyst (Table 2). Several one-pot prepared supports were used and numerous experiments were performed for each supports in ordered to investigate their performance and verify the reproducibility of the process. The data reported in Table 2 (Entry 1, 3, 5 and 7) are average values for several metathesis experiments conducted for similar catalysts supported over ZnCl2-OMA supports prepared with the same aluminium precursor and carboxylic acid.
Entry | Catalysta | Conv.b (%) | Metathesis productsc (%) | DMPd (%) | Se | Ref. | ||||
---|---|---|---|---|---|---|---|---|---|---|
(1) trans | (1) cis | (2) trans | (2) cis | (3) | ||||||
a The investigated 3 wt% MTO/ZnCl2-OMA catalysts supports were prepared using different aluminium precursors and carboxylic acids, while meso-Al2O3 represents the wormhole-like mesoporous alumina.b Conv.: conversion, expressed as Conv. (%) = ([methyl oleate]0 − [methyl oleate]t)/[methyl oleate]0 × 100.c The numbers appearing for the metathesis products 1, 2, and 3 (Scheme 1) are expressed in terms of individual product yield (yieldi), defined as: yieldi (%) = [producti]/([methyl oleate]0 − [methyl oleate]t) × 100, where [producti] is the molar concentration of the product i = 1 to 3, and [methyl oleate]0 and [methyl oleate]t are the molar concentrations of methyl oleate initially and at a given time t, respectively.d DMP: desired metathesis products (1 + 2) yield, which is the summation of the individuals yields of 1 and 2.e The selectivity (S) is defined as the yield of all desired products over the yields of all undesired products, which represents; S = yield (1 and 2)/yield of (3). op: ZnCl2-OMA supports prepared through one-pot process. ts: ZnCl2-OMA supports prepared through two-steps process. | ||||||||||
1 | 3 wt% MTO/ZnCl2-OMA-1 (Al(OBus)3 + citric, op) | 80.2 | 13.7 | 12.7 | 14.9 | 6.9 | 32.0 | 48.2 | 1.51 | This work |
2 | 3 wt% MTO/ZnCl2-OMA-1 (Al(OBus)3 + citric, ts) | 78.2 | 15.5 | 14.3 | 9.1 | 4.3 | 35.0 | 43.2 | 1.23 | Abidli et al.28 |
3 | 3 wt% MTO/ZnCl2-OMA-2 (Al(OBus)3 + tartaric, op) | 93.6 | 48.8 | 17.1 | 5.4 | 2.4 | 19.9 | 73.7 | 3.70 | This work |
4 | 3 wt% MTO/ZnCl2-OMA-2 (Al(OBus)3 + tartaric, ts) | 88.2 | 29.3 | 8.5 | 12.9 | 4.6 | 32.9 | 55.3 | 1.68 | Abidli et al.28 |
5 | 3 wt% MTO/ZnCl2-OMA-3 (Al(OPri)3 + tartaric, op) | 86.7 | 34.8 | 9.4 | 9.1 | 2.5 | 30.9 | 55.8 | 1.81 | This work |
6 | 3 wt% MTO/ZnCl2-OMA-3 (Al(OPri)3 + tartaric, ts) | 87.4 | 31.8 | 8.9 | 14.1 | 2.4 | 30.2 | 57.2 | 1.89 | Abidli et al.28 |
7 | 3 wt% MTO/ZnCl2-OMA-4 (Al(OBut)3 + tartaric, op) | 79.8 | 14.1 | 13.2 | 13.6 | 6.4 | 32.5 | 47.3 | 1.46 | This work |
8 | 3 wt% MTO/ZnCl2–meso-Al2O3 (ts) | 86.7 | 24.4 | 8.3 | 16.2 | 3.6 | 34.3 | 52.4 | 1.53 | Pillai et al.25–27 |
9 | 3 wt% MTO/meso-Al2O3 | 2.4 | 0.2 | 0.3 | 0.2 | 0.3 | 1.4 | 1.0 | 0.71 | Pillai et al.25–27 |
10 | 2nd generation Grubbs catalyst | 89.6 | 20.1 | 3.9 | 21.0 | 4.2 | 40.4 | 49.2 | 1.22 | Pillai et al.25–27 |
The other alumina supports used for comparison include the zinc chloride-promoted (Table 2, Entry 8) and unpromoted (Table 2, Entry 9) conventional wormhole-like alumina that we have reported earlier;29–31 as well as the well-ordered hexagonal ZnCl2-OMA prepared via two steps process that we have also recently reported (Table 2, Entry 2, 4 and 6).32 In addition, the newly one-pot prepared ZnCl2-OMA-based heterogeneous metathesis catalyst performance (Table 2, Entry 1, 3, 5 and 7) has been compared to that of the homogeneous commercially available and widely used 2nd generation Grubbs catalyst (Table 2, Entry 10).
Briefly, the addition of zinc chloride to alumina supports was found to offer a great catalytic promoting effect for the metathesis reaction.29–31 The unmodified wormhole-like alumina supports showed the weakest metathesis performance among all tested catalysts, while their promoted counterparts showed much more enhanced catalytic performance (Table 2, Entry 8).29–31 Afterwards, we recently showed that the zinc chloride-promoted hexagonally organized mesoporous alumina (ZnCl2-OMA) materials are the most suitable support of MTO catalysts showing relatively better methyl oleate conversions.32 In all cases methyl oleate self-metathesis leads to the formation of both desired metathesis products including 9-octadecene and dimethyl-9-octadecene-1,18-dioate (Scheme 2) as well as undesired product (methyl elaidate, Scheme 2). However, using the ordered alumina supports, improved selectivity was reached towards the desired metathesis products (Table 2, Entry 2, 4 and 6). More interestingly, we demonstrated that the new catalyst design based on organized mesoporous network has an enormous effect on improving the reaction rate for the conversion of such bulky molecule (methyl oleate).32 This enhancement is attributed to the absence of diffusion limitations of substrates and metathesis products within the ordered cylindrical mesopores of these supports. Compared to the disordered and interconnected pores of the wormhole-like mesoporous alumina supports where molecular diffusion limitations and blocking effects may occurs. All these features were also provided using similar one-pot synthesized ZnCl2-OMA supports, which offered also better activity than the MTO-based catalysts supported over zinc chloride-modified and unmodified wormhole-like alumina (Table 2).
The catalytic performance of the successfully prepared ZnCl2-OMA supports via a one-pot synthesis process studied in this work was also compared to the performance of the ZnCl2-OMA-based catalysts previously prepared through two steps process. Interestingly the one-pot prepared ZnCl2-OMA materials were found to be the best catalytic support for MTO. Metathesis reaction data displayed in Table 2 (Entry 1, 3, 5 and 7) shows higher catalytic performance including; methyl oleate conversion (up to 93.6%), desired metathesis products yield (up to 73.7%) and selectivity (up to 3.7), using these catalytic supports. For the structure-activity correlation, we believe that this enhanced performance may arise from various factors.
Firstly, the newly one pot-prepared ZnCl2-OMA supports exhibits larger BET surface area (Table 1) compared to the two steps-prepared ZnCl2-OMA supports. The lower BET surface area of these later is due to the subsequent calcination step and pore filling with zinc chloride upon post-modification, which alters their textural properties. The BET specific surface area was dramatically dropped upon incorporation of zinc chloride.32 On the other hand, during the one pot synthesis of ZnCl2-OMA supports this additional annealing step was avoided with in situ impregnation of zinc chloride. This is in a good agreement with several studies showing that supported catalysts' activity is often proportional to the specific surface area of these catalysts.76–83 Furthermore, higher BET surface area favors better dispersion of supported MTO species, which improves the concentration and availability of metathesis active species and intermediates allowing more efficient use of exposed supported MTO molecules. Secondly, unlike the two steps synthesis process, avoiding a second calcination step during the one-pot preparation process will reduce the dehydroxylation of the alumina surface as illustrated above in XPS and MAS NMR sections. Along with this, we showed that zinc chloride is incorporated on the alumina via interactions with surface hydroxyl groups. As a result, inhibited dehydroxylation may allow better incorporation of Zn active intermediates which are crucial for the MTO metathesis catalytic action, displaying a key synergetic effect. Finally, we believe that the single-step approach allow simultaneous zinc species deposition and mesopores walls formation, releasing more homogeneous and well-dispersed zinc species on the OMA surface. Moreover, this simultaneous process may result in a confinement effect of ordered mesochannels of OMA inhibiting Zn species aggregation and growth. This agglomeration phenomenon may likely occur during the two steps synthesis which restricts Zn species interactions with alumina surface and limits their availability for the catalytic process. It is also worth to mention that the separate wet impregnation method during the two-steps synthesis was found to alter the textural and morphological properties of the OMA (mechano-chemical effect), along with formation of zinc chloride aggregates, as revealed by SEM images.32
These results are very exciting showing higher performance, even better than those obtained with the traditional highly active 2nd generation Grubbs catalyst (Table 2, Entry 10), with slightly better methyl oleate conversion (93.6% vs. 89.6%), but exhibiting much more selectivity towards desired metathesis products (73.7% vs. 49.2%). It is worth to mention that we previously showed that the OMA-supported MTO catalyst was found to exhibit fast kinetics,32 comparable to that observed over the Grubbs catalyst,31 with the possibility of recycling and reuse for further metathesis runs.
In addition to the enhanced catalytic activity of the MTO-based metathesis catalysts using the newly prepared ZnCl2-OMA materials, this one-pot synthesis process presents further advantages. Economically, compared to the two-steps process, this process allowed a reduced use of solvents. Also, it reduces energy consumption through elimination of additional evaporation and calcination steps, while reducing enormously the preparation time. More interestingly, using the one-pot process we reached higher synthesis yields, obtaining higher amounts of ZnCl2-OMA materials, almost 75% higher in weight than those obtained using the two steps process. This was mainly attributed to limitation of weight loss during the two subsequent calcination processes.
Footnote |
† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra12057h |
This journal is © The Royal Society of Chemistry 2015 |