Highly dispersed copper (oxide) nanoparticles prepared on SBA-15 partially occluded with the P123 surfactant: toward the design of active hydrogenation catalysts

B. Dragoi a, I. Mazilu ab, A. Chirieac a, C. Ciotonea abc, A. Ungureanu *a, E. Marceau c, E. Dumitriu a and S. Royer *bc
a“Gheorghe Asachi” Technical University of Iasi, Faculty of Chemical Engineering and Environmental Protection, 73 D. Mangeron Bvd, 700050 Iasi, Romania. E-mail: aungureanu@tuiasi.ro
bUniversité de Poitiers, CNRS UMR 7285, Institut de Chimie des Milieux et Matériaux de Poitiers (IC2MP), Bâtiment B35, 6 Rue Michel Brunet – TSA 51106, 86073 Poitiers Cedex 9, France. E-mail: sebastien.royer@univ-lille1.fr
cUniv. Lille, CNRS, ENSCL, Centrale Lille, Univ. Artois, UMR 8181 – UCCS – Unité de Catalyse et de Chimie du Solide, F-59000 Lille, France

Received 19th May 2017 , Accepted 23rd July 2017

First published on 24th July 2017


Copper (oxide) nanoparticles with dispersion and size ranging from highly dispersed nanoparticles confined within the intra-wall pores to nanoparticles confined within the main mesopores can be simply produced via incipient wetness impregnation by using mesoporous silica supports containing the structure directing agent P123 partially occluding the intra-wall pores of SBA-15. Remarkably high copper dispersions of 69.5 down to 25.9% for metal loadings from 5 to 20 wt% were achieved as a result of the dual effect of the presence of surface silanol groups and the unique confining space between the silica walls and residual P123 in the intra-wall pores. The copper catalysts demonstrated outstanding activities in the hydrogenation of cinnamaldehyde, with TOF values ranging from 5.1 × 10−3 to 22.1 × 10−3 s−1, which are much higher than those of other supported copper catalysts reported elsewhere and prepared using conventional approaches.


Introduction

The benefits brought by supported nanoparticles (NPs) in various fields, such as energy conversion and storage, chemical manufacturing, chemical sensing, medical diagnostics, green production of fuels and chemicals, nanoelectronics and so on,1–6 are generally recognized throughout the chemistry community, and represent the driving force of continuous interest in the design and preparation of NPs with controlled compositions, sizes, and morphostructures. Non-noble metal (oxide) NPs, as efficient nanocatalysts in many chemical reactions, have attracted special attention in the last few years because they represent a feasible alternative to the less abundant and expensive noble-metal-based NPs.7 Copper-based catalytic materials with a high dispersion of NPs (i.e., below 10 nm in size) are of particular interest because copper shows unique catalytic properties in various processes, such as oxidation,8 hydrogenation,9 cycloaddition10 and so on. However, the preparation of copper NPs of such size is challenging because of the mobility of copper species that makes them difficult to be stabilized on the surface of a solid support such as silica. Usually, supported CuO or Cu particles with sizes of 15–20 nm or even higher are obtained by conventional impregnation and drying, followed by high-temperature air calcination, and reduction.11 Different synthetic methods,12–18 controlled conditions of calcination/reduction thermal treatments,19 formation of solid solutions by addition of a second metal,12,20 and use of various supporting materials (ordered mesoporous materials, layered double hydroxides, phyllosilicates, etc.),21–23 are among the most straightforward strategies to increase the thermo-chemical stability of copper-based NPs, a property that determines their long-term size and dispersion, as well as catalytic properties. Nonetheless, the experimental results indicated that the stability of Cu NPs is far from being resolved, especially at high metal loading and by applying impregnation as the catalyst preparation method. For instance, copper loaded on supports with different chemical surfaces (γ-Al2O3, SiO2, CeO2, Fe2O3) and pore organization (SiO2, MCM-48) by impregnation and using acethylacetonates as precursors and a copper loading between 3–8 wt% generated CuO NPs of sizes included in the range 6.2–21.3 nm: γ-Al2O3 – 6.2 nm, SiO2 – 18.5 nm, CeO2 – 9.5 nm, Fe2O3 − 21.3 nm, MCM-48 – 21.3 nm.21 Particle size in the range between 2–10 nm for metallic copper, with a metal loading of 10 wt%, was obtained when copper nitrate was loaded on SiO2 and SBA-15 mesoporous silica by the impregnation method, but only applying carefully controlled calcination conditions (NO or N2) and using Zn as a promoter.13 In another example, SBA-15-supported copper catalysts with Cu/Si up to 20% containing highly dispersed copper particles (1–4 nm in size) were prepared by impregnation followed by special vacuum–thermal treatments.23

Among the various support materials proposed to stabilize Cu NPs, SBA-15 mesoporous silica with a dual pore system, micropores and mesopores, appeared as a versatile support offering a multitude of properties to be explored. Particularly, the chemistry of the surface is very important and it can be adjusted by different strategies in view of the targeted properties. For the specific case of template-mediated porous supports, the chemical state of the surface is significantly influenced by the approach applied to remove the structure directing agent, the P123 surfactant in the case of SBA-15 silica, from the pores. Hence, the support surface exhibits only silanol groups when P123 is removed by calcination (combustion). Due to the high temperature applied to fully eliminate P123, the framework suffers contraction as a result of the surface dehydroxylation and formation of siloxane bonds, thus diminishing the concentration of silanols,24,25 which are very important anchoring sites for the metallic precursors. On the other hand, following the ethanol extraction method, SBA-15 still contains P123 inside the intra-wall pores, while the concentration of silanols remains higher in comparison with the solid subjected to calcination.25 In addition, the accessibility of metal precursors to the internal surface should be facilitated due to the partial exclusion of the P123 surfactant from the main mesopores.

Herein, we demonstrate that SBA-15 containing residual P123 inside the pores, obtained by ethanol extraction, is an effective support material able to stabilize copper (oxide) phases in the absence of promoters as highly dispersed NPs with controlled morphostructures through a simple impregnation procedure, without special calcination/reduction treatments. Based on our previous results,26–29 it was believed that the generated confined space between the hydrophilic PEO groups of P123 and silica walls provides a suitable environment to stabilize NPs at low sizes, by limiting the mobility of the metal precursors during thermal treatment steps of calcination and reduction. The approach presented herein is very convenient and represents a sustainable alternative to other sophisticated methods involving carefully controlled synthesis conditions and the use of water sensitive organic precursors and organic solvents.

Experimental

Chemicals

All chemicals required to prepare the materials were used without any additional purification: tetraethylorthosilicate (Si(OC2H5)4, TEOS, 98%, Aldrich), non-ionic triblock co-polymer Pluronic P123 (poly(ethylene oxide)-block-poly(propylene oxide)-block-poly(ethylene oxide), EO20PO70EO20, molecular weight = 5800, BASF Corp.), copper nitrate (Cu(NO3)2·3H2O, 98%, Aldrich), distilled water and hydrochloric acid. The chemicals used for the hydrogenation reaction were also used as received: trans-cinnamaldehyde (C6H5–CH[double bond, length as m-dash]CH–CHO, 98%, Merck) as a reagent and iso-propanol (C3H8O, 99%, Sigma-Aldrich) as solvent.

Preparation of samples

SBA-15 support. SBA-15 was synthesized according to a classical procedure, as proposed by Zhao et al.30 4 g of Pluronic P123 and 1.6 M HCl solution were stirred at 40 °C until the complete dissolution of the surfactant. Then, 8.5 g of TEOS was added dropwise under magnetic stirring and the mixture was aged for 24 h. The resulting gel was transferred into a polypropylene bottle and heated at 100 °C for 48 h. After filtration and drying, half of the support was subjected to calcination at 550 °C for 6 h in a muffle oven (heating ramp of 1.5 °C min−1) to remove the organic template, while the other part was subjected to ethanol extraction for 5 h in order to retain only ∼15% of P123 inside the pores (content confirmed by TG analysis, Fig. S1).
Coppers catalysts. CuO containing SBA-15 materials with different loading degrees of metallic copper (5, 10, and 20 wt%) were prepared by incipient wetness impregnation, followed by mild drying (IWI-MD). Freshly calcined and partially extracted SBA-15 supports were impregnated with aqueous solutions of the corresponding hydrated nitrates to obtain the desired concentration of metal. The resulting materials were dried at 25 °C for 5 days. The copper oxide phases were obtained after calcination under stagnant air at 500 °C for 6 h (heating ramp of 1.5 °C min−1). Samples were denoted as Cux/SBA-15_y were, x = 5, 10, and 20 wt%, whereas y = c – calcined and e – extracted, respectively.

Physico-chemical characterisation

Thermogravimetric (TG) measurements. The thermogravimetric (TG) measurements were performed using a computer-coupled Q-derivatograph (MOM). This analysis was performed under stagnant air in the temperature range 25–600 °C (heating rate of 10 °C min−1) for the supports with different contents of the P123 surfactant. Likewise, the supports were analysed by FTIR spectroscopy using a Scimitar FTS 2000 (Digilab) apparatus, at a spectral resolution of 4 cm−1 on KBr wafers.
Inductively coupled plasma optical emission spectrometry. ICP-OES was performed on a Perkin sequential scanning spectrometer to determine the elemental composition of the catalysts (Cu and Si). Before analysis, a known amount of calcined sample was introduced in a diluted HF–HCl solution and then digested under microwave radiation.
Powder X-ray diffraction. XRD analysis was performed on a Bruker AXS D5005 X-ray diffractometer, using Cu Kα radiation (λ = 1.54184 Å) as the X-ray source. For low-angle analysis, the data were collected in the 2θ range from 0.75 to 5° with a step of 0.01° (step time of 10 s). For high-angle analysis, the data were collected in the 2θ range from 10 to 80° with a step of 0.05° (step time of 8 s). Crystal phase identification was made by comparison with the ICDD database.
Nitrogen physisorption. N2 physisorption experiments were carried out on an Autosorb 1-MP automated gas sorption system (Quantachrome Instruments). Prior to analysis, the samples were outgassed under high vacuum at 350 °C for 3 h. The adsorption/desorption isotherms were obtained at −196 °C allowing 4 min for equilibration between successive points. Textural properties were determined from the isotherms by using the Autosorb 1 software, version 1.55. The Brunauer–Emmet–Teller method was applied to obtain the surface area (SBET) and non-local density functional theory (NL-DFT) for cylindrical pores was applied to evaluate the pore size distribution (PSD). The pore volume was evaluated on the adsorption branch of the isotherm at p/p0 = 0.97.
High resolution transmission electronic microscopy coupled with energy dispersive X-ray spectroscopy. TEM-EDXS was used to characterize the pore structure of the SBA-15 support, distribution of NPs throughout the pores, and morphostructure of NPs. The micrographs were obtained on a JEOL 2100 instrument (operated at 200 kV with a LaB6 source and equipped with a Gatan Ultra scan camera). EDXS was carried out with a Hypernine (Premium) detector (active area: 30 mm2) using the software SM-JED 2300T for data acquisition and treatment. The EDXS analysis zone is defined on the particle, and generally ranges from 5 to 15 nm. Before analysis, the sample was first included in a resin and then a cut of ∼100 nm width was realized by ultramicrotomy.
Temperature programmed reduction (H2-TPR) and N2O chemisorption. Experiments were performed on a ChemBET Pulsar TPR/TPD from Quantachrome. For the reducibility experiments, about 30 mg of calcined sample was inserted in a U-shaped microreactor. Before each TPR run, the catalyst was activated at 500 °C for 1 h under a flow of simulated air (40 mL min−1). After cooling to 50 °C, the H2 containing flow was stabilized (40 mL min−1, 5 vol% H2 in Ar) and TPR was performed from 50 to 500 °C with a temperature ramp of 5 °C min−1. The metallic surface area (SCu) was determined by the nitrous oxide chemisorption method. In a typical experiment, about 30 mg sample was placed in a U-shaped quartz reactor and subjected to activation and reduction under the same conditions as those used for TPR experiments and kept for 2 h at 500 °C (TPR1). Then, the sample was cooled to 60 ± 5 °C under Ar flow and then it was exposed to a N2O flow for 0.5 h (40 mL min−1). The accessible Cu0 surface species were oxidized to Cu2O by N2O. The sample was again flushed with pure Ar for 0.5 h, and then a second TPR run (the same conditions as those for TPR1) was performed in order to reduce Cu2O to Cu0 (TPR2). The copper dispersion was calculated by using the following formula: D (%) = (2A2/A1) × 100, where A2 and A1 stand for the area of the reduction peaks obtained by TPR2 and TPR1, respectively.

Hydrogenation of cinnamaldehyde

For each test, the calcined catalysts were reduced under hydrogen flow (1 L h−1) at 350 °C for 10 h (heating rate of 6 °C min−1). The catalytic tests were carried out in a high-pressure Parr reactor under the following conditions: 1 mL of aldehyde, 40 mL of isopropanol, 0.265 g of catalyst, a hydrogen pressure of 10 bar, and a reaction temperature of 130 °C. Samples were periodically taken from reactor and analysed by GC with an HP 5890 series gas chromatograph, equipped with a DB-5 capillary column and a flame ionization detector. The identification of the reaction products was achieved from the retention times of pure compounds and occasionally by GC-MS (Agilent 6890N system equipped with an Agilent 5973 MSD detector and a DB-5-ms column). The conversion of cinnamaldehyde and selectivities to the different hydrogenation products were calculated by taking into account the FID response factors for each compound.

Results and discussion

X-ray diffraction

Fig. 1 (left) displays the XRD patterns at high-angle for the copper containing samples. It can be noticed that support treatment from calcination to extraction significantly affects the properties of the copper phase formed, which is always CuO (tenorite; ICDD 048−1548).31 Hence, for the sample prepared on the calcined SBA-15 support, at a copper loading of 5 wt% (Cu5/SBA-15_c), intense diffraction peaks corresponding to the CuO crystalline phase can be observed. By applying the Scherrer equation, a mean crystallite size of 24.7 nm was calculated for this sample (Table 2). It is obvious that some of the copper precursors segregated at the external surface of the support grains, which can explain why a particle size larger than the SBA-15 pore size is obtained (see Fig. S3). This indicates the instability of the highly mobile copper precursors on the silica surface of the calcined SBA-15, with a limited density of Si–OH surface groups acting as anchoring sites for metallic precursors,32 which drastically decreased upon calcination25 to liberate silica porosity. Results obtained over the calcined SBA-15 show therefore that the corresponding surface is clearly not adequate to effectively anchor the copper species and produce dispersed NPs.
image file: c7cy01015j-f1.tif
Fig. 1 High-angle (left) and low-angle (right) XRD patterns of the CuO/SBA-15 solids.

When 5 wt% copper was loaded on SBA-15 containing residual P123 organic surfactant (Cu5/SBA-15_e), the obtained XRD pattern was completely different. Only the peak at ∼23° attributed to amorphous silica can be identified in the diffractogram. When comparing these samples, i.e. those prepared on calcined and extracted SBA-15, respectively, the amount of copper precursor cannot be taken into consideration to explain the absence of the diffraction peaks in the diffractogram of Cu5/SBA-15_e. Therefore, the lack of diffraction peaks in the diffractogram of the Cu5/SBA-15_e sample can be explained by the obtained small-size particles of CuO.

Such formation of small particles can be rationalized by the higher density of surface silanol groups and the presence of residual organic surfactant inside the intra-wall pores, which drastically diminish the mobility of copper species during thermal treatment. Indeed, we showed previously that copper can be stabilized as very small and highly dispersed NPs, when non-calcined solids containing organic structure directing agents are used as supports.26,27 It was possible to load up to 35 wt% copper on such supports, but only using a solvent free impregnation route, i.e. melt infiltration. In another study, the stabilization of copper (5 wt% of metal with a Cu[thin space (1/6-em)]:[thin space (1/6-em)]Ni mass ratio of 4[thin space (1/6-em)]:[thin space (1/6-em)]1) was achieved on a partially extracted SBA-15, but in the presence of nickel as a second metal.28 Similarly, there are reports in the literature on the dispersion of mono-component copper NPs on mesoporous materials by solvent free procedures. For instance, Wang et al. reported a loading of up to 20 wt% copper by simply physically mixing copper precursors with calcined SBA-15, followed by controlled heating.33 Quite recently, ∼3–7 wt% of copper was dispersed by two-step high-energy ball milling and in situ pyrolysis reduction procedures on a mesoporous phenol-formaldehyde polymer.34 However, in these examples, the migration of the metal species was avoided by eliminating the solvent for copper phase preparation. Indeed, the presence of solvent is the origin of the redistribution of the precursor into the porous structure prior to its decomposition, with formation of large aggregates inside the pores or even their migration out of pores. Hence, it is the first report on copper oxide loaded by solvent impregnation on SBA-15 with partially occluded intra-wall pores, with formation of small – XRD not detectable – particles distributed inside an ordered silica support. It is therefore clear that the organic functionality of the surfactant, i.e., the hydrophilic ethylene oxide groups (EO), played an essential role in facilitating the retention of copper species on the surface and their distribution either into the intra-wall pores in the space formed between the silica walls and EO groups, or on the internal surface of the mesopores in the form of small NPs not detected by XRD.

In order to validate this hypothesis, i.e., the presence of residual P123 and Si–OH groups that simultaneously contribute to the stabilization of copper NPs, TG and FT-IR experiments were performed on the calcined, partially extracted and as-made SBA-15 silica supports. The recorded curves are included in Fig. S1 and S2. Firstly, the TG results (Fig. S1) clearly indicate that residual P123 was retained in the support subjected to ethanol extraction for 5 h, with a weight loss of ∼15 wt% being registered during thermal degradation under air. In contrast, the organic residues are not found in the calcined SBA-15 support, with the weight loss in the interval 100–600 °C being only ∼3 wt% which could be attributed to the minor surface dehydroxylation processes. IR results (Fig. S2) are in excellent agreement with the TG data. Hence, for the partially extracted and as-made support materials, vibration bands between 2800–3100 and 1300–1500 cm−1 were observed for both samples, which are assigned to the C–H stretching and bending vibrations, respectively, of the surfactant P123.24 For the calcined SBA-15, these bands are not seen in the corresponding spectrum, demonstrating the complete removal of the organic template during the thermal process. It should be however emphasized that the amount of the organic molecule in the extracted sample is lower in comparison with that in the as-made sample. This is in line with the TG data, because, only a part of the organic template was removed by the extraction process. Another important band attributed to the bending vibration of silanol groups is identified at 960 cm−1.24 It is remarkable that the intensity of this band can be correlated with the treatment applied to liberate the pores, which also influences the amount of the remaining Si–OH groups in the sample. Hence, for the calcined sample, the band at 960 cm−1 is hardly detectable, while for the partially extracted and as-made samples, these bands are very well shaped and of similar intensities. This finding indicates that both samples contain essentially similar amounts of silanols. It is thus obvious that the thermal treatment drastically changes the surface chemistry of the support by transforming the silanol groups into siloxane ones, while a softer and eco-friendly treatment, such as ethanol extraction, does not change the chemical nature of the as-synthesized mesoporous silica, by keeping the number of Si–OH residues constant. Previous quantitative studies showed that the concentration of silanols in extracted SBA-15 supports is much higher in comparison with the solids subjected to calcination (i.e., 8.5 vs. 3.4 OH per nm2).25 Therefore, our experimental results provide direct evidence of the synchronised contribution of Si–OH and residual P123 in anchoring copper based NPs loaded on such interesting and promising hybrid mesoporous supports.

By increasing the metal loading to 10 wt% (Cu10/SBA-15_e), the high-angle XRD pattern exhibits the characteristic peaks of the CuO phase. The shape of the peaks, which are very broad and not very intense, suggests the growth of the crystalline phases mainly inside the mesopores, as confined NPs. If the NPs are confined, then the mean particles size cannot significantly surpass the mesopore diameter, i.e., ∼8–9 nm. The mean crystallite size calculated with the Scherrer equation is 9.5 nm (Table 2). Interestingly, when the copper loading is further increased to 20 wt% (Cu20/SBA-15_e), the width of the peaks remains large while the intensity increases. The mean crystallite size calculated by decomposing the massif in the 30–40° zone in two components remains similar to Cu10/SBA-15_e (i.e., 9.9 vs. 9.5 nm; Table 2). On the other hand, the ratio between the intensities of the peaks corresponding to the planes (002) and (200) (I(002)/I(200)) is 2.55 for this sample, which is much higher than the corresponding ratios for the samples with lower copper content (i.e., 1.01 for Cu5/SBA-15_c; 1.37 for Cu10/SBA-15_e), and much higher than the conventional value (0.4).35 Although it is quite difficult to assign the crystal growth in a certain direction and thus, the morphology of the crystals from the intensity of the XRD peaks, an assumption on it could be done.36,37 It was previously shown that the change in growth direction and morphology of nanocrystals can alter the intensity of the XRD peaks.38 On the basis of this assumption combined with the profile of the peaks, it appears that the shape of the crystals grown in the confining space between silica walls and EO groups is changed as the amount of copper precursors infiltrated in such a small space increases. Hence, it can be supposed that the resulted crystals are preferentially oriented so that the {100} facets are abundant.35 Therefore, it can be stated that the chemical properties of the support and the amount of copper precursor could be used to enhance the control over the size, distribution and morphostructure of the produced copper oxide NPs.

The small-angle XRD pattern of the partially extracted SBA-15 support (SBA-15_e) (Fig. 1 right) displays one intense peak at 0.86° corresponding to the (100) plane and two additional peaks at 1.4 and 1.6° attributed to the (110) and (200) planes, respectively, indicating an ordered 2D hexagonal mesostructure. The corresponding d-spacing of the (100) plane is 10.2 nm and the a0 parameter is 11.8 nm (Table 1). After deposition of the copper precursors, drying and calcination, the registered diffractograms exhibit comparable reflections, demonstrating the preservation of the hexagonal mesostructure in the CuO-containing materials. The (100) peak shifted to slightly higher 2θ values after copper loading, which suggests an increase of the pore thickness probably as a result of copper deposition on the internal surface of the pores. It is well documented that the reflection of the (100) plane is generated by the electron density contrast between the empty pores and the silica walls.39 As the contrast is higher, the intensity of this reflection is higher. Interestingly, the intensity of the (100) peak seems to be not affected after copper deposition, suggesting that most of the NPs found on the internal surface are small enough to have a limited effect on the electron density contrast between the empty space and silica walls. Instead, as the loading of copper increases, the ratio between the intensities of the diffraction peaks at higher 2θ i.e., I110/I200, becomes gradually higher, increasing in the following order: Cu5/SBA-15_e < Cu10/SBA-15_e < Cu20/SBA-15_e (the corresponding values are: 1.09, 1.10, 1.23). This result could be explained by the progressive localization of the NPs inside the intra-wall pores that become filled, thus reducing the contrast between the intra-wall pores and silica walls.

Table 1 Chemical composition, textural and structural properties of the supports and copper-based samples
Sample Cu,a wt% d 100,b nm a 0,c nm S BET,d m2 g−1 V p,e cm3 g−1 D p,f nm
a Cu loading obtained by ICP-OES. b d 100 = the lattice spacing obtained by low angle XRD. c a 0 = the hexagonal unit cell parameter obtained by low angle XRD image file: c7cy01015j-t1.tif. d S BET = the total specific surface area by the BET equation. e V p = the total pore volume at p/p0 = 0.97. f D p = the average pore size evaluated by NL-DFT for cylindrical pores/equilibrium model.
SBA-15 c 9.1 10.5 882 1.21 8.5
SBA-15_e 10.2 11.8 802 1.86 9.0
Cu5/SBA-15_e 5.4 9.7 11.2 638 1.10 8.4
Cu10/SBA-15_e 10.5 9.9 11.4 689 1.03 6.0; 8.4
Cu20/SBA-15_e 20.7 9.9 11.4 643 0.95 7.0


N2 physisorption

The textural properties of the samples were determined by nitrogen physisorption at −196 °C. The corresponding isotherms and pore size distribution curves (PSD) are illustrated in Fig. 2. The main textural parameters are listed in Table 1. The support prepared by ethanol extraction (SBA-15_e) displays an isotherm of type IV with a hysteresis of type H1 according to the IUPAC classification, reflecting a well ordered mesostructure, in agreement with the small-angle XRD results. The single step capillary condensation at p/p0 = 0.75 confirms the presence of cylindrical pores with a narrow distribution of size, centred at 9.0 nm. The BET surface area of this sample is 802 m2 g−1, while the pores provide a volume of 1.86 cm3 g−1. For the calcined support, the BET surface area is 882 m2 g−1, which is 10% higher than that of SBA-15_e (802 m2 g−1). For the total pore volume, the SBA-15_e has a larger pore volume (1.86 cm3 g−1) than SBA-15_c (1.21 cm3 g−1), which can be explained by the less shrinkage of the silica framework of SBA-15_e during the extraction with ethanol.25 As shown in Table 1, the values of these parameters generally decrease for the samples containing CuO NPs because these particles are located inside the pores of the support.
image file: c7cy01015j-f2.tif
Fig. 2 N2 physisorption isotherms (left) and pore size distribution curves (right) for the CuO containing materials and used SBA-15_e support.

Capillary condensation takes place at almost the same relative pressures as for the support and only a small shift at lower p/p0 can be noticed (from 0.75 to 0.72 for Cu5/SBA-15_e and Cu10/SBA-15_e and 0.70 for Cu20/SBA-15_e). This indicates a small decrease in the diameter of the mesopores due to the localization of the NPs on their surface. Indeed, the maxima of the PSD curves for these samples are observed at 8.4 nm for Cu5/SBA-15_e and Cu10/SBA-15_e and 7.0 nm for Cu20/SBA-15_e.

A particular situation arose for the Cu10/SBA-15_e sample for which the desorption branch contains a large step before the closing point, which is reflected in the PSD curve as a new maximum centred at 6.0 nm. Such a shape of the desorption branch indicates the formation of confined NPs inside the mesopores that generate the so-called ink bottle-like pores responsible for the second maximum in the PSD curve.40 Thus, for the sample containing 10 wt% copper, the presence of highly dispersed and mesopore confined NPs can be envisaged, which is in well agreement with the high-angle XRD results discussed above. As the amount of copper further increases (Cu20/SBA-15_e), the pore blockage became more evident, the adsorption/desorption branches of the isotherms losing the parallelism between each other. The pore size distribution is broader suggesting a wide range of pores with different sizes whose average is at 7.0 nm.

Microscopy

Cux/SBA-15_e materials were further analysed by TEM. Representative images at low and high magnification are displayed in Fig. 3.
image file: c7cy01015j-f3.tif
Fig. 3 TEM images for the oxide forms of copper based materials with different degrees of copper loading.

In the low magnification TEM image of the sample Cu5/SBA-15_e, no bulk particles can be distinguished. However, when high magnification was applied, very small dark spots appeared, but only after exposure to the electron beam for at least 3 min, which increases the contrast between silica and NPs, due to the possible reduction of copper oxide under the beam. Thus, after exposure for 1 min, no particles can be detected (Fig. S4), which suggests that they are embedded within the silica walls, i.e. NPs are confined within the intra-wall pores. As it can be observed in Fig. 3 and S4, the analysed zone is not stable under the electron beam, and exposition times longer than 5 min lead to an alteration of the mesostructure, though a better visualization of the copper nanoparticles can be made. For this material, the particle size distribution curve, determined from micrograph statistical analysis, after prolonged exposure, indicates particles with size in the range 0.5–3 nm, with an average size of ∼1.7 nm (Fig. 4).


image file: c7cy01015j-f4.tif
Fig. 4 Particle size distributions evaluated by TEM for the calcined copper based materials.

For the Cu10/SBA-15_e sample, the low magnification TEM images reveal the presence of small NPs homogeneously distributed throughout the support, whereas the high resolution images reveal sphere-like NPs with a diameter very close or slightly smaller to that of the mesopores. It should be mentioned that such a morphology of CuO nanoparticles is difficult to obtain by simple methods, in turn sophisticated SBA-15-surface functionalization schemes with in situ reduction should be applied.41 The particle size distribution, obtained by treatment of the micrographs (Fig. 4), reflects the size evolution of the observed particles, ranging from 4 to 15 nm, and a mean value of 9.1 nm, very close of the pore size of the support.

For Cu20/SBA-15_e, most of the NPs were always observed inside the porous structure of the support, demonstrating that even at a high concentration, the copper phase can be mostly introduced inside the channels when SBA-15_e is used as an organic–inorganic hybrid scaffold for impregnation. However, the obtained NPs' distribution throughout the silica grains was not as homogeneous as those for the other two samples. Hence, the TEM images reveal that some of the copper migrated in zones of the support where the concentration was higher and where silica walls broke during calcination (Fig. 3, zone A). Therefore, NPs with size in the range 10–30 nm were formed. The images taken for the second kind of zone (Fig. 3, zone B) display an ordered mesostructure. Noteworthily, in this area the NPs are located along the mesopores surface and in the silica walls, and they show an average size ranging from 3 to 10 nm. Consequently, a bimodal particle size distribution can be considered for this sample (Fig. 4, Table 2).

Table 2 Crystallite (XRD) and particle (TEM) sizes of CuO for all samples, as well as N2O chemisorption data
Sample D CuO, nm N2O chemisorption
XRDa TEMb D,c % d Cu,d nm S Cu,e m2Cu gcat−1
a The crystallite size of CuO evaluated by the Scherrer equation (dhkl = λ/β[thin space (1/6-em)]cos[thin space (1/6-em)]θ). b The particle size of CuO determined from the TEM images. c Dispersion degree of metallic copper calculated from N2O chemisorption. d Average size of Cu particles. e Cu exposed surface area.
Cu5/SBA-15_c 24.7 30–100 4.6 21.6 1.4
Cu5/SBA-15_e 1.7 69.5 1.4 25.4
Cu10/SBA-15_e 9.5 9.1 44.6 2.2 31.7
Cu20/SBA-15_e 9.9 ∼5; 15 25.9 3.8 36.3


TPR and N2O chemisorption

The reducible properties of the oxide forms of the catalysts as well as the metal–support interactions were investigated by TPR. The corresponding TPR profiles are depicted in Fig. 5.
image file: c7cy01015j-f5.tif
Fig. 5 TPR profiles for the calcined copper based materials.

The TPR curve of the sample prepared on the calcined support (Cu5/SBA-15_c) exhibits one reduction maximum centered at 285 °C, corresponding to the reduction of Cu2+ to Cu0 in bulk-like, large CuO crystallites.12,20 It can be thus supposed that the metal–support interaction is weak for the Cu5/SBA-15_c sample.

For the samples prepared on SBA-15 with pores partially occluded by P123, on-set temperatures at ∼190 °C can be observed in the corresponding TPR profiles, indicating the higher reducibilities of these materials as compared to Cu5/SBA-15_c. The reduction temperature of CuO depends strongly on the particle size and dispersion, with the reducibility being increased as the dispersion increases.42–44

The TPR profiles of Cu5/SBA-15_e and Cu10/SBA-15_e samples are relatively sharp, with a small shoulder appearing at low temperature. The observed reduction maxima are located at ∼250 °C and ∼260 °C, respectively, in line with a homogeneous size distribution of the highly-dispersed CuO NPs, as already shown by TEM. In contrast, the TPR profile of the Cu20/SBA-15_e material becomes slightly broader, with the reduction maximum at ∼255 °C, suggesting a more heterogeneous size distribution of the CuO NPs, as also concluded from Fig. 4. However, no visible reduction shoulder at ∼285 °C attributed to bulk-like CuO particles is observed, confirming that these particles are mostly located inside the pores of SBA-15.

The copper exposed surface area, percentage of dispersion and Cu particles size were obtained by the N2O dissociative chemisorption method. This method involves the dissociation of N2O over accessible metallic copper sites, according to the reaction:

N2O + 2Cu(s) → N2 + Cu2O(s)

The values obtained are listed in Table 2. It can be observed that the lowest dispersion (4.6%) was obtained for the copper loaded on SBA-15_c, while the highest dispersion (69.5%) was measured for the copper loaded on the partially occluded mesopores of SBA-15_e, at the same copper loading (5 wt%). For these two samples, copper particles sizes of 21.6 and 1.4 nm, respectively, were calculated. It is clearly indicated that the maintaining of residual P123 inside the pores can generate an adequate environment for the production of small, highly dispersed NPs throughout the pores of the support. The dispersion degree decreases as the amount of copper increases in the sample, when using SBA-15_e. The dispersion degree still remains at high values: 44.6% and 25.9% for Cu10/SBA-15_e and Cu20/SBA-15_e, respectively. These results confirm interpretations of data derived from TEM and XRD analyses. The values obtained here by this simple approach are in line with our previous studies on copper loaded on polyether-functionalized ordered mesoporous silica for which similar dispersions were reported.27

Catalytic results

Cinnamaldehyde (CNA) is a molecule containing two double bonds in the side carbon chain, i.e. olefin (C[double bond, length as m-dash]C) and aldehyde (C[double bond, length as m-dash]O) bonds. As illustrated in Scheme 1, the hydrogenation of CNA takes place via parallel and consecutive reactions when partially hydrogenated (hydrocinnamaldehyde – HCNA and cinnamyl alcohol – CNOL) and totally hydrogenated products (hydrocinnamyl alcohol – HCNOL) are obtained.
image file: c7cy01015j-s1.tif
Scheme 1 Reaction pathways for the hydrogenation of cinnamaldehyde.

The evolution of conversion of CNA with reaction time is displayed in Fig. 6 for all produced catalysts. As a first observation, the catalyst prepared on the calcined support (Cu5/SBA-15_c) manifested extremely low activity, with the conversion of CNA after 360 min being only 8.4 mol%. This result is not surprising, considering the low activity of copper when in the form of large particles.27 For the catalyst prepared using the support containing residual P123 (Cu5/SBA-15_e), at the same loading, a significantly higher activity is obtained, with a total conversion of CNA reached in 180 min. Obviously, increasing the amount of copper in the sample, while maintaining high dispersion (Table 2), generated catalysts with much higher activity. Thus, the catalyst containing 10 wt% copper completely converted the CNA molecule in 60 min, whereas for the catalyst with 20 wt% copper, only 30 min was necessary to totally convert CNA.


image file: c7cy01015j-f6.tif
Fig. 6 CNA conversion vs. reaction time obtained over the copper catalysts (test conditions: Treduction = 350 °C; Treaction = 130 °C, 0.265 g catalyst; 1 mL of CNA, 40 mL iso-propanol as solvent, PH2 = 10 bar, stirring rate = 700 rpm).

The efficiency of the catalysts investigated herein was also evaluated as reaction rates, and as turnover frequency (TOF, expressed in moles of CNA converted per mole of surface copper per time unit). The values are listed in Table 3.

Table 3 Selectivity to reaction products at ∼40 mol% of CNA conversion, initial rates and TOF values for all catalysts
Sample Selectivity, mol% r 0, μmol gcat−1 s−1 TOF × 103, s−1
HCNA CNOL HCNOL
Cu5/SBA-15_c 0.06 1.7
Cu5/SBA-15_e 48.9 28.0 23.1 3.13 5.1
Cu10/SBA-15_e 59.6 14.2 26.2 8.43 10.9
Cu20/SBA-15_e 43.0 11.2 45.8 19.53 22.1


As expected on the basis of N2O chemisorption results, the reaction rate increases with the copper loading in Cux/SBA-15_e catalysts. For the divided materials, TOF is also observed to significantly evolve with the copper dispersion in the materials. Thus, as observed in Table 3, TOF increases by a factor of about 4, from 5.1 × 10−3 s−1 (Cu5/SBA-15_e) to 22.1 × 10−3 s−1 (Cu20/SBA-15_e), before significantly decreasing down to 1.7 × 10−3 s−1 over poorly dispersed Cu5/SBA-15_c. Obviously, the performance of the catalysts is enhanced as the amount of copper increases in the samples, however, TOF evolution shows significant modification of the intrinsic activities of the surface copper site with particle size. Then, a too high dispersion (as in Cu5/SBA-15_e) results in a lower intrinsic activity of sites than that obtained over slightly larger particles (as in Cu20/SBA-15_e), while external large particles show very low intrinsic activity (Cu5/SBA-15_c). In line with previous investigations, these results can be tentatively explained by the structure-sensitivity of cinnamaldehyde hydrogenation reaction over divided Cu catalysts, since support effects can be precluded.45–47 Accordingly, in the case of highly dispersed Cu5/SBA-15_e, the activity of more unsaturated copper surface sites is likely to be inhibited by stronger adsorption of reactants, whereas in the case of less dispersed Cu20/SBA-15_e, the cinnamaldehyde conversion probably takes places predominantly on more densely packed copper surface sites where the adsorption of reactants is expected to be less strong. The comparison of TOF values for copper catalysts with those from the literature indicates that the copper catalysts synthesized in this work are among the most active supported copper-based catalysts. Thus, under similar reaction conditions, a TOF of 3 × 10−3 s−1 was obtained for 12 wt% copper loaded on ordinary SiO2, whereas 2 × 10−3 s−1 was calculated for 21 wt% copper loaded on MCM-41 mesoporous silica.21 These two values are not far from the low TOF value recorded for Cu5/SBA-15_c. Once again, these results highlight the great advantage brought by the preservation of a small amount of hydrophilic organic residues of the P123 surfactant inside the pores of the support to stabilize copper as very small NPs that are homogeneously dispersed on the internal surface of the mesoporous silica, while loading is a simple way to adjust the dispersion up to an adequate size.

The selectivity values obtained for these catalysts (Table 3) are in the classical range of values for copper, with the highest selectivity for CNOL being noted for the catalyst with the lowest amount of copper and the smallest size, i.e., 28% for a conversion of ∼40%. Interestingly, the selectivity to CNOL decreases as the amount of copper increases in the sample, while the contribution of HCNOL increases. The distribution of the reaction products as a function of conversion is plotted in Fig. S5 for the active catalysts. The product distribution curves show that at the early stage of reaction, the catalysts are selective to CNOL or HCNA (the partially hydrogenated products), depending on the copper loading in Cux/SBA-15_e, and with the increase in conversion the selectivity to these products decreases, while the selectivity to HCNOL (the totally hydrogenated product) increases. These trends indicate the presence of two different adsorption sites on the copper NP surface on which the CNA molecule will adsorb via either C[double bond, length as m-dash]C or C[double bond, length as m-dash]O bonds that are hydrogenated in parallel. Thus, as the reaction proceeds, a part of CNA is constantly converted to HCNA, whereas the rest of it is transformed in CNOL; both HCNA and CNOL are further hydrogenated to HCNOL. As the amount of copper increases (and copper particle size increases), the trends of selectivity are modified, with HCNOL being generated by hydrogenation of HCNA instead of CNOL. It clearly indicates a preferential adsorption of CNA via the C[double bond, length as m-dash]C bond. Interestingly, for the catalyst with 10 wt% copper, the change of the selectivity trend (transformation of HCNA into HCNOL) takes place at high conversion, i.e., ∼75%, while for the catalyst with 20 wt% copper, this transformation is observed at a conversion of 40%. These results might underline the role of the crystalline plane of the active phase exposed to the reactants, in agreement with theoretical and experimental studies reported by different research groups on Pt showing that the CNA molecule is adsorbed via the C[double bond, length as m-dash]O bond on Pt(111) planes, while the adsorption occurs via both double bonds on the (100) plane.48–50 It is thus expected that a similar situation occurs in the case of copper especially since crystallites can grow so that the (100) planes become more exposed as the amount of copper increases.51 The crystal orientation during synthesis appears to affect the selectivity by modifying the adsorption process of CNA molecules, and directing preferentially the hydrogenation through C[double bond, length as m-dash]C or C[double bond, length as m-dash]O bonds.

Conclusions

Our results revealed that the combined effects of the silanols and confining space between EO hydrophilic groups of the residual P123 and silica walls of SBA-15 make this support ideal to control the growth of copper as highly dispersed NPs. Furthermore, the NP size and localization can be modified by loading to promote the formation of mesopore confined NPs instead of highly divided intra-wall confined NPs. The use of SBA-15 containing residual hydrophilic groups is consequently efficient to control the size and localization of the Cu nanoparticles that could be extended to other mono- or bi-component transition metal (oxide) NPs loaded on mesoporous silica, such as those based on Ni, Co, Fe, Mn, etc. Moreover, these small NPs encapsulated in the pores of SBA-15 exhibit particularly high catalytic activity in the hydrogenation of cinnamaldehyde. High TOF values are obtained, confirming the potential of copper as an effective hydrogenation catalyst, when present in a divided state.

Acknowledgements

This work was partially supported by two grants from the Romanian National Authority for Scientific Research, CNCS-UEFISCDI (project numbers PN-II-RU-TE-2012-3-0403 and PN-II-ID-PCE-2011-3-0868). C Ciotonea (784016 L) and I. Mazilu (812773E) acknowledge the Ministère des affaires étrangères et du développement international for Eiffel Excellence fellowships. C. Ciotonea acknowledges the Fondation de l'Université de Poitiers for financial support. S. Royer, E. Marceau and C. Ciotonea acknowledge Chevreul institute (FR 2638), Ministère de l'Enseignement Supérieur et de la Recherche, Région Nord – Pas de Calais and FEDER for funding.

Notes and references

  1. M. E. Franke, T. J. Koplin and U. Simon, Small, 2006, 2, 36 CrossRef CAS PubMed .
  2. R. Wilson, Chem. Soc. Rev., 2008, 37, 2028 RSC .
  3. Y. Roman-Leshkov, C. J. Barret, Z. Y. Liu and J. A. Dumesic, Nature, 2007, 447, 982 CrossRef CAS PubMed .
  4. H. M. Torres Galvis, J. H. Bitter, C. B. Khare, M. Ruitenbeek, A. Iulian Dugulan and K. P. de Jong, Science, 2012, 335, 835 CrossRef CAS PubMed .
  5. G. Schmid, Chem. Soc. Rev., 2008, 37, 1909 RSC .
  6. A. S. Arico, P. Bruce, B. Scrosati, J. M. Tarascon and W. Van Schalkwijk, Nat. Mater., 2005, 4, 366 CrossRef CAS PubMed .
  7. F. Zaera, Chem. Soc. Rev., 2013, 42, 2746 RSC .
  8. M.-J. Suh and S.-K. Ihm, Top. Catal., 2010, 53, 447 CrossRef CAS .
  9. K. Yoshida, C. Gonzalez-Arellano, R. Luque and P. L. Gai, Appl. Catal., A, 2010, 379, 38 CrossRef CAS .
  10. A. Shaygan Nia, S. Rana, D. Döhler, F. Jirsa, A. Meister, L. Guadagno, E. Koslowski, M. Bron and W. H. Binder, Chem. – Eur. J., 2015, 21, 10763 CrossRef CAS PubMed .
  11. V. Gutierrez, M. Dennehy, A. Diez and M. A. Volpe, Appl. Catal., A, 2012, 437–438, 72 CrossRef CAS .
  12. A. Ungureanu, B. Dragoi, A. Chirieac, C. Ciotonea, S. Royer, D. Duprez, A. S. Mamede and E. Dumitriu, ACS Appl. Mater. Interfaces, 2013, 5, 3010 CAS .
  13. G. Prieto, J. Zečević, H. Friedrich, K. P. de Jong and P. E. de Jongh, Nat. Mater., 2013, 12, 34 CrossRef CAS PubMed .
  14. T. Tsoncheva, A. Gallo, I. Genova, I. Spassova, M. Marelli, M. Dimitrov, M. Khristova, G. Atanasova, D. Kovacheva, D. Nihtyanova and V. Dal Santo, Inorg. Chim. Acta, 2014, 423, 145 CrossRef CAS .
  15. K. Yoshida, C. Gonzalez-Arellano, R. Luque and P. L. Gai, Appl. Catal., A, 2010, 379, 38 CrossRef CAS .
  16. S. Wang, W. Guo, H. Wang, L. Zhu, S. Yin and K. Qiu, New J. Chem., 2014, 38, 2792 RSC .
  17. C. Ciotonea, B. Dragoi, A. Ungureanu, A. Chirieac, S. Petit, S. Royer and E. Dumitriu, Chem. Commun., 2013, 49, 7665 RSC .
  18. C.-S. Chen, C.-C. Chen, C.-T. Chen and H.-M. Kao, Chem. Commun., 2011, 47, 2288 RSC .
  19. P. Munnik, M. Wolters, A. Gabrielsson, S. D. Pollington, G. Headdock, J. H. Bitter, P. E. de Jongh and K. P. de Jong, J. Phys. Chem. C, 2011, 115, 14698 CAS .
  20. B. Dragoi, A. Ungureanu, A. Chirieac, V. Hulea, S. Royer and E. Dumitriu, Catal. Sci. Technol., 2013, 3, 2319 CAS .
  21. V. Gutierrez, M. Alvarez and M. A. Volpe, Appl. Catal., A, 2012, 413–414, 358 CrossRef CAS .
  22. B. Dragoi, A. Ungureanu, A. Chirieac, C. Ciotonea, C. Rudolf, S. Royer and E. Dumitriu, Appl. Catal., A, 2015, 504, 92 CrossRef CAS .
  23. C.-H. Liu, N.-C. Lai, J.-F. Lee, C.-S. Chen and C.-M. Yang, J. Catal., 2014, 316, 231 CrossRef CAS .
  24. B. Tian, X. Liu, C. Yu, F. Gao, Q. Luo, S. Xie, B. Tu and D. Zhao, Chem. Commun., 2002, 1186 RSC .
  25. L. Wang and R. T. Yang, J. Phys. Chem. C, 2011, 115, 21264 CAS .
  26. C. Rudolf, I. Mazilu, A. Chirieac, B. Dragoi, F. Abi-Ghaida, A. Ungureanu, A. Mehdi and E. Dumitriu, Environ. Eng. Manage. J., 2015, 14, 399 CAS .
  27. C. Rudolf, F. Abi-Ghaida, B. Dragoi, A. Ungureanu, A. Mehdi and E. Dumitriu, Catal. Sci. Technol., 2015, 5, 3735 CAS .
  28. A. Chirieac, B. Dragoi, A. Ungureanu, C. Ciotonea, I. Mazilu, S. Royer, A. Sophie Mamede, E. Rombi, I. Ferino and E. Dumitriu, J. Catal., 2016, 339, 270 CrossRef CAS .
  29. C. Ciotonea, I. Mazilu, B. Dragoi, C. Catrinescu, E. Dumitriu, A. Ungureanu, H. Alamdari, S. Petit and S. Royer, ChemNanoMat, 2017, 2, 233 CrossRef .
  30. D. Zhao, J. Feng, Q. Huo, N. Melosh, G. H. Fredrickson, B. F. Chmelka and G. D. Stucky, Science, 1998, 279, 548 CrossRef CAS PubMed .
  31. http://database.iem.ac.ru/mincryst/s_carta.php?TENORITE  .
  32. Nanoscale Materials in Chemistry, ed. K. Klabunde and R. Ryans, John Wiley and Sons, Inc., 2nd edn, 2009 Search PubMed .
  33. Y. M. Wang, Z. Y. Wu and J. H. Zhu, J. Solid State Chem., 2004, 177, 3815 CrossRef CAS .
  34. K. Wang, L. Yang, W. Zhao, L. Cao, Z. Sun and F. Zhang, Green Chem., 2017, 19, 1949 RSC .
  35. Y. Sun and Y. Xia, Science, 2002, 298, 2176 CrossRef CAS PubMed .
  36. D. P. Singh, A. K. Ojha and O. N. Srivastava, J. Phys. Chem. C, 2009, 113, 3409 CAS .
  37. Z. Fei, P. Lu, X. Feng, B. Sun and W. Ji, Catal. Sci. Technol., 2012, 2, 1705 CAS .
  38. B. G. Ganga, M. R. Varma and P. N. Santhosh, CrystEngComm, 2015, 17, 7086 RSC .
  39. M. Imperor-Clerc, D. Bazin, M. D. Appay, P. Beaunier and A. Davidson, Chem. Mater., 2004, 6, 1813 CrossRef .
  40. J. R. A. Sietsma, J. D. Meeldijk, M. Versluijs-Helder, A. Broersma, A. Jos van Dillen, P. E. De Jong and K. P. De Jong, Chem. Mater., 2008, 20, 2921 CrossRef CAS .
  41. V. S. Garcia-Cuello, L. Giraldo and J. C. Moreno-Pirajan, J. Chem. Eng. Data, 2011, 56, 1167 CrossRef CAS .
  42. C. H. Tu, A. Q. Wang, M. Y. Zheng, X. D. Wang and T. Zhang, Appl. Catal., A, 2006, 297, 40 CrossRef CAS .
  43. X. C. Zheng, S. H. Wu, S. P. Wang, S. R. Wang, S. M. Zhang and W. P. Huang, Appl. Catal., A, 2005, 283, 217 CrossRef CAS .
  44. Z. Liu, M. D. Amiridis and Y. J. Chen, J. Phys. Chem. B, 2005, 109, 1251 CrossRef CAS PubMed .
  45. A. J. Plomp, H. Vuori, A. O. I. Krause, K. P. de Jong and J. H. Bitter, Appl. Catal., A, 2008, 351, 9 CrossRef CAS .
  46. Y. Zhu and F. Zaera, Catal. Sci. Technol., 2014, 4, 955 CAS .
  47. R. van den Berg, G. Prieto, G. Korpershoek, L. I. van der Wal, A. J. van Bunningen, S. Lægsgaard-Jørgensen, P. E. de Jongh and K. P. de Jong, Nat. Commun., 2016, 7, 13057 CrossRef CAS PubMed .
  48. F. Delbecq and P. Sautet, J. Catal., 1995, 152, 217 CrossRef CAS .
  49. V. Ponec, Appl. Catal., A, 1997, 149, 27 CrossRef CAS .
  50. Y. Yuan, S. Yao, M. Wang, S. Lou and N. Yan, Curr. Org. Chem., 2013, 17, 400 CrossRef CAS .
  51. Y. Zheng, J. Liang, Y. Chen and Z. Liu, RSC Adv., 2014, 4, 41683 RSC .

Footnote

Electronic supplementary information (ESI) available: Supplementary characterization results. See DOI: 10.1039/c7cy01015j

This journal is © The Royal Society of Chemistry 2017