Template-induced in situ dispersion of enhanced basic-sites on sponge-like mesoporous silica and its improved catalytic property

Abstract

Catalytic performance of heterogeneous catalysis is strongly dependent on the dispersity of catalytic active sites, and especially a high exposure of the unit active phase is promising for the overall catalytic process. In this study, a novel strategy was developed to fabricate an unprecedented CaO-based mesoporous solid strong base catalyst. Relying on the physicochemical assembly of Ca²⁺ in the interface between a micelle and siliceous wall, thin-layer-like calcium oxide species were formed in situ and dispersed in the mesochannels of silica. Wherein, the gradual coverage of CaO on the mesoporous wall was controlled by adjusting the amounts of Ca²⁺ on alkylamine micelles. Interestingly, a novel sponge-like microscale structure of silica was discovered in the CaO-based mesoporous-composites for the first time, which completely differs from the reported mesoporous silica. More importantly, the introducing of a CaO solid base on the pore wall is nearly non-destructive for the textural properties of the mesoporous matrix. The direct template-induced mesoporous solid strong base not only received extremely dispersed and unexpected enhanced strong basic sites (CO₂-desorption temperature ≥718 °C), but also avoided repeated thermal processes for the degradation of the basic resource, and saved energy and time. This heterogeneous alkaline catalyst shows excellent catalytic activity for the synthesis of dimethyl carbonate under a milder reaction condition (30 °C, 25 min) and holds stability and reusability beyond comparison with the conventional catalysts. The dispersed and enhanced strong basic sites, combined with excellent mesoporous properties, are demonstrated to be responsible for such a high catalytic performance.

---

1. Introduction

Biodiesel, as sustainable energy, has been gradually applied in practical production due to the consumption of limited fossil fuels.^1–3 Wherein, the usage of a highly efficient heterogeneous alkaline catalyst in the transesterification of biodiesels is an intriguing research orientation for obtaining many valuable chemicals from some biological sources as a substitution for limited fossil fuels.^4–6 Recently, dimethyl carbonate (DMC) as an eco-friendly chemical is now widely accepted due to its negligible toxicity and efficient utilization of carbon dioxide in the synthesis process.

Generally, heterogeneous base catalysis is regarded as the interface reaction process. Therefore, the direct fabrication of highly active heterogeneous solid base catalysts with a high exposure of unit catalytic active phase has been demonstrated to be promising due to the rapidly increasing accessibility of the active sites. Many endeavors have been continuously devoted to increasing the dispersity and surface exposure degree of active species to improve their catalytic activity.^7–10 Mesoporous solid bases have attracted great interest due to some individual advantages, including improved performance, simple separation, no corrosion, and negligible waste production.^11–13 So far, increasing attention has been paid to developing a highly-efficient mesoporous solid base catalyst possessing a high surface area and well-dispersed strong basic sites.^6,14–18

Directly fabricating a mesoporous solid base using basic metal sources as the mesoporous matrix is not fully reasonable due to the unavailability of massive unexposed basic sites in the framework. In addition, the obtained mesoporous solid base also failed to give the desirable structure properties such as large surface area.^19,20 Hence, many post-treatment methods have been developed by incorporating alkaline sources on the supports with a large surface area, such as mesoporous silica, to obtain some preferable porous solid bases. For instance, through grafting special organic bases or anchoring nitrogen-bearing species, some interesting Lewis basic sites can be generated on the pore surface of mesoporous silica.^14,21–23 Despite all this, the generated basicity on these materials is still relatively weak as compared to some counterparts owing to the inherent nature of basic species. Fortunately, the decoration of strong basic metal oxides (e.g. K₂O, CaO and MgO) on the surface of a mesopore has also been demonstrated to be efficient for enhancing the base strength.^10,24–27 Typically, the modification of basic metal oxides for mesoporous silica mainly involves the introduction of basic metal sources in mesoporous silica and followed by calcination to produce the basic sites on the pore surface of supports. However, noticeably, the introduction of basic metal oxides renders a drastic decrease of surface area and pore size of supports by the conventional method. In addition, the introduction of basic metal oxides in the mesochannel is also uncontrollable and nonuniform, and generally accompanied with the growth and sintering of large basic oxide particles, hence failing to produce a desirable solid base with a full decoration of the mesoporous surface. More apparently, energy consumption in the repeated calcination, including removing of the micelle template of the porous material and decomposition of basic metal salts is very high. Consequently, only weak basicity is obtained in the resulting materials, accompanied with serious damage with regard to the mesostructure of the supports. Despite these efforts, the development of an efficient method to create strong basic sites on mesoporous silica remains a great challenge to date.

Herein, based on our previous works, we report a simple templating method by directly introducing basic metal Ca²⁺ into the interface of the micelle and pore wall during the assembly process and then in situ producing well-dispersed and high exposure enhanced CaO active basic species by a thermal process. Wherein, Ca²⁺ not only acted as the base source, but also played a co-structure directing role in the self-assembly process. The metallomicelles formed from the coordination of Ca²⁺ and the DDA (dodecylamine) micelle served as a cation-like micelles template, which facilitates improvement of the structure of the matrix due to the presence of stronger countra-ion interaction between the metallomicelles and siliceous oligomers. As a result, thin-layer-like CaO strong basic species were formed in situ and dispersed on the mesoporous wall of silica during the calcined process. The synthetic strategy is schematically illustrated in Scheme 1. The involved catalysts were characterized using various efficient techniques to demonstrate the status of CaO and the structure of the material. Finally, the obtained solid base catalysts were applied for the synthesis of dimethyl carbonate to evaluate their catalytic performance, and the special basic sites were studied for this reaction. The catalytic reaction mechanism of dimethyl carbonate synthesis over the mesoporous solid base was proposed and clarified.

Scheme 1 Template-induced in situ dispersion of solid basic sites on mesoporous silica (A) as well as the conventional thermal method (B).

2. Experimental

2.1. Materials

Dodecyl amine (DDA, ≥98%) was obtained from Aladdin. Calcium acetate monohydrate (Ca(CH₃COO)₂·H₂O, ≥99%), tetraethylorthosilicate (TEOS, AR) and methanol (MeOH, ≥99%) were purchased from Sinopharm Chemical Reagent Co., Ltd. Propylene carbonate (PC) was purchased from Aladdin. All chemicals were commercially available and used without any purification.

2.2. Preparation of xCaO–HMS (hexagonal mesoporous silica)

The synthetic process of the Ca²⁺-participated assembly CaO–HMS was operated as follows: DDA (1.85 g, 0.01 mol) was blended with ethanol (25 mL) and deionized water (62.5 mL), and then the mixture was stirred at 45 °C for 30 min. Controllable contents of Ca(Ac)₂·H₂O (1, 2, 3 4 and 5 mmol) were added and resulted in the status of solutions turning from milky to translucent. After being restored to room temperature, TEOS (10.73 mL) was added dropwise under vigorous stirring for another 4 h and subsequently kept static at 45 °C for 18 h. In the end, the resulting samples were collected by centrifugation and washed with water and ethanol three times. The as-prepared samples were calcined at 550 °C for 6 h in dry air steam. The final catalysts corresponding to gradually increasing calcium contents were labelled as xCaO–HMS (x = 1, 2, 3, 4, 5), and their theoretical molar ratio of Ca/Si is 2.08, 4.17, 6.24, 8.32 and 10.4 (%). The comparative catalyst was prepared based on a typical wet post-impregnated method involving the introduction of a calcium source in bare HMS followed by calcination, and then the resulting catalyst was designated as 5CaO–HMS(p). The molar ratio of Ca/Si was controlled at about 10.4%.

2.3. Characterizations

The XRD patterns of all of the samples were collected using Smartlab TM 9 KW (Rigaku Corporation, Tokyo, Japan) equipped with a rotating anode and Cu Kα radiation (λ = 0.154178 nm).

The N₂ adsorption–desorption tests were carried out using BELSORP-MINI volumetric adsorption analyzer (BEL Japan, Osaka, Japan). The resulting samples were outgassed in vacuum at 150 °C for 3 h and before measurements. The specific surface areas and pore size distribution were calculated through Brunauer–Emmett–Teller (BET) and Barrett–Joyner–Halenda (BJH) methods, respectively.

The basicities of the samples were examined by CO₂-temperature programmed desorption. The samples were pretreated under a He stream at 200 °C for 1 h (100 mL min⁻¹). Subsequently, the temperature of system was cooled down to 100 °C, and a flow of pure CO₂ (50 mL min⁻¹) was subsequently introduced into the reactor. Desorption of CO₂ was performed from 100 to 800 °C (10 °C min⁻¹) under a He flow (30 mL min⁻¹), and CO₂ was detected with TCD signals.

Fourier transform infrared (FT-IR) spectra of the samples were collected on a Bruker VECTOR22 spectrometer (Bruker, Bruker, Germany).

Scanning electron microscopy (SEM) images were obtained from a Hitachi S4800 Field Emission Scanning Electron Microscope (Hitachi, Tokyo, Japan). The elemental distribution of materials was determined with energy dispersive X-ray spectroscopy (EDX).

High-resolution transmission electron microscopy (HRTEM) images were collected on a JEM-2010 EX microscope (JEOL, Tokyo, Japan) equipped with EDX.

The X-ray photoelectron spectra (XPS) were recorded on a PHI 5000 Versa Probe X-ray photoelectron spectrometer (ULVAC-PHI, Kanagawa, Japan) equipped with Al K_α radiation (1486.6 eV). The binding energy of C 1s at 284.6 eV was utilized as the reference.

The calcium contents in the catalysts were examined using a PE Optima 2000DV Inductively Coupling Plasma emission spectrometer (ICP) (PerkinElmer, Waltham, MA, USA). All the samples were pretreated and completely dissolved with hydrofluoric acid.

Computational details. All the calculation results were based on the Gaussian 09 package. Density of atom charge (Hirshfeld) was calculated with the Becke's three-parameter hybrid exchange functional and the Lee–Yang–Parr correlation functional (B3LYP).

2.4. Catalysis assessment

The catalytic performances of the series of xCaO–HMS were assessed using the transesterification reaction of propylene carbonate and methanol. Before the reaction, all the CaO-based HMS materials were activated at 800 °C in a pure N₂ flow for 1 h. The reaction was carried out in a 25 mL flask under strong magnetic stirring at a certain temperature in the desirable time. The molar ratio of MeOH to PC in all the test reactions was 4 : 1, the used catalyst amount was 0.50 wt% of MeOH. Finally, the product was examined on a SP-6890 gas chromatograph (Lunan Ruihong Chemical Instrument Co. Ltd, China) with a 0.32 mm × 30 m SE-54 capillary column after centrifugal separation. The conversion and selectivity were calculated by an external standard method.

In order to investigate the special effect of the γ basic sites in the synthesis of DMC, we conducted a designed experiment with CO₂ absorption and desorption treatments at a specific temperature (550 °C) to cover the γ basic sites and obtain CaO–HMS containing only α and β basic sites (see Scheme 4). For the detailed operation refer to the above CO₂-TPD characterization section. The CO₂ treated CaO-based HMS were further used to catalyze the synthesis of DMC under the reaction conditions as same as that (50 °C, 60 min) of 5CaO–HMS(p).

Finally, the stability of the catalysts was examined by the recycling experiments under their respective optimized reaction conditions.

3. Results and discussion

3.1. Synthetic clarification

The synthetic procedure for the in situ dispersion of CaO on the mesoporous wall of silica is illustrated in Scheme 1. It was proposed that the self-assembly mechanism of the material is through a Ca²⁺-participating templating process based on our previous works.^28,29 As we know, the traditional HMS materials prepared by the neutral templating route (S⁰I⁰) are based on hydrogen bonding and self-assembly interaction between neutral primary amine surfactants (S⁰) and neutral inorganic precursors (I⁰). Herein, the introduced Ca²⁺ cations served as important intermediaries in this templating process setting up a bridge between the DDA micelle template and the siliceous species by a coordinated and electrostatic interaction. In this process, Ca²⁺ cations first coordinated on dodecyl amine to generate a metallomicelle (Ca²⁺–DDA) which is similar to a cationic-like micelle template. Massive cationic Ca²⁺ were enriched on the surface of DDA micelles. Therefore, negatively-charged silicate oligomers from the hydrolysis of TEOS are easily assembled on the surface of Ca²⁺ modified DDA micelles by a countra-ion intermediary interaction. As a result, this mesoporous HMS solid base decorated with well-dispersed CaO species on the surface would be generated in situ after calcining to remove the template.

3.2. Characterizations for the special CaO-based solid base catalyst

Fig. 1a displays the low-angle XRD patterns of the series xCaO–HMS. As shown, all the samples give a diffraction peak at 2–3° which is attributed to the (100) diffraction peak of the mesophase. The presence of the (100) feature peak demonstrates the mesostructure of all the samples. In addition, an obvious shift to a lower angle for the (100) diffraction of samples from 1CaO–HMS to 5CaO–HMS can be observed, which indicates the enlargement of d₁₀₀ lattice spacing. This should be due to the increment of CaO basic species on the pore wall with the incorporation of more Ca²⁺ in the assembly process.²⁸ Interestingly, with the gradual introducing of Ca²⁺ on the DDA micelles, the intensity of mesophase diffraction exhibits a gradual rise and subsequent drop process, and the maximum of intensity is reached with regard to the sample of 2CaO–HMS. This result suggests a variation of the mesoporous regularity of the series xCaO–HMS with different CaO loadings, which is associated with the participation of Ca²⁺ in the assembly process. The coordination of Ca²⁺ on the DDA micelles results in the enrichment of positive charges on the micelle template and the resulting metallomicelles would more easily match the negative-charge of Si–O oligomers in the assembly process. The enhanced countra-ion assembly interaction promotes the formation of an improved structure. However, excessive gathering of Ca²⁺ on micelles also may destroy the pristine micelles' structure due to a mutual repulsion of positive charges on the micelles.²⁸ Therefore, the introduction of more Ca²⁺ leads to a rational drop of mesoporous regularity.

Fig. 1 Low-angle (a) and wide-angle (b) XRD patterns of various samples.

Fig. 1b shows the high-angle XRD patterns of all the samples. Each sample shows a broad peak at around 20–30°, which is attributed to amorphous silica. However, no crystal diffraction peak attributed to CaO is detected in all the curves of the series xCaO–HMS synthesized from the templating route, which indicates that the Ca species introduced may be uniformly dispersed on the pore wall and do not aggregate into a bulk CaO crystal. This result favors the formation of a thin-layer-like dispersed solid base on the mesoporous wall, which is also in accordance with TEM results (Fig. 3). In contrast, it is obviously observed that three strong diffraction peaks present in the curves of post-impregnated sample 5CaO–HMS(p) are typically attributed to cubic CaO crystals,²⁶ which demonstrates the production of large CaO crystals in post-impregnated 5CaO–HMS(p). Associated with the assembly mechanism of the series of CaO–HMS materials, this suggests that the formation of extremely-dispersed CaO should be due to the strong physicochemical interaction between the metal calcium and the siliceous pore wall, which efficiently inhibits the migration and aggregation of CaO during the high-temperature thermal treatment process.

Fig. 2 Representative SEM images of series samples (a) 1CaO–HMS, (b) 2CaO–HMS, (c) 3CaO–HMS, (d) 4CaO–HMS, (e) 5CaO–HMS and (f) photo of a sponge.

Fig. 3 Representative TEM images (a–d) and its elemental mapping images (e) of sponge-like 5CaO–HMS, the white scale bar represents 2 μm.

Fig. 1S† presents the nitrogen sorption results and pore size distribution of the series xCaO–HMS. As exhibited in Fig. 1Sa,† the N₂-adsorption/desorption isotherms of all the samples exhibit a IV type curve and H1 hysteresis loop, indicating the typical mesostructured materials. In addition, it should be noted that all samples show high adsorption capacity (ca. 2.1 cm³ g⁻¹) especially in the high relative pressure region of 0.9–1.0. The sudden jump at a high relative pressure indicates the presence of massive stacked voids. This result reveals that the xCaO–HMS may be stacked with fine-sized particles, and many voids exist between them. Further detailed textural parameters calculated from nitrogen desorption data are summarized in Table 1S†. All the samples exhibit a similar pore size (ca. 2.6 nm) and high specific surface area (ca. 1000 m² g⁻¹), which manifests a fact that the incorporation of Ca²⁺ in the assembly process shows little effect on the pore diameter and specific surface area. However, in comparison with the post-synthesized 3CaO/HMS(p) sample, all the CaO–HMS catalysts prepared from the metal-assisted template method show obviously superior textural properties, which highlights the superiority of this self-assembly process. In contrast, this is very difficult to reach in the traditional method by introducing metal species followed by a calcining process, which is generally accompanied with a large detriment to the structure of the matrix material. Note that the series xCaO–HMS samples even exhibit a higher structural characteristic as compared to the pure HMS. This result manifests the introducing Ca²⁺ in the assembly process to obtain a high surface area of solid base catalysts is very promising. It is suggested that the presence of a better structure should be attributed to the improved matching interaction by intermediary countra-ions during the assembly process of mesoporous silica, which is supported by reported work.²⁸ Another characteristic worth noting is that the calculated pore wall thickness of the series xCaO–HMS exhibits a rising trend with increments of CaO in the mesopore. Apparently, the increasing pore wall thickness herein should be attributed to the CaO species dispersed in situ on the pore wall.

Fig. 2 displays a representative morphology and the partial characteristics of series samples for xCaO–HMS from scanning electron microscopy. The low-power SEM images show that the whole morphology of series xCaO–HMS consists of large spherical particles and some broken fractions. Whereas, the higher-power SEM images show that the surface of these particles is in a fluffy status and seems to be heaped up by ultra-fine nanoparticles, with massive voids present in the sphere just like a sponge structure, which is a good explanation for the presence of a high adsorption capacity. This evidence is also in good accordance with N₂-adsorption/desorption results that there is the presence of a high jump of adsorption volume at high relative pressure. In addition, as compared to the status of the microsphere in pure HMS samples (Fig. 2S†), an obvious difference can be found, and the surface fluffy status was absent in pure HMS. It can be concluded that the participation of Ca²⁺ in the assembly process of HMS has some special effects on the morphology.

The morphology and structural features of materials can be further examined by transmission electron microscopy. Fig. 3 shows representative TEM images and elemental mapping images of sample 5CaO–HMS. Distinctly, the 5CaO–HMS shows a special microscale structure (multi-void sponge-like characteristic), which has not as yet been reported for the mesoporous silica materials. Typically, this material consists of massive nanoscale particles, and the porosity of the particles is also observed based on the images of (a) and (b), which is similar to the reported wormhole-like pore structure.^30,31 It should be noted that this structural characteristic is universal in the overall material and not only as a partial characteristic. In addition, the observed results are in good agreement with the SEM surface characteristics. This is also a reasonable explanation for the presence of the high pressure jump of the adsorption/desorption curve in N₂-adsorption/desorption isotherms. On the other hand, the TEM results reveal that the visual CaO crystalline species is not observed; however, the EDX spectrum demonstrates the existence of the Ca element, suggesting that the dispersed CaO species should be incorporated in the mesochannel of silica in the form of an amorphous status. Meanwhile, elemental mapping results also confirm that the resulting material is comprised of Si, O and Ca, and their uniform distribution in the selected region of mesocomposite indicates the homogeneous dispersity of CaO on the surface of the silica. Finally, with respect to the formation of multi-particle stacked CaO-based silica, a surface-energy-dominated process is regarded to be the result. The generation of metallomicelles during the assembly process may result in mutual repulsion among the micelles and thereby make the nucleated particles assembled from the metallomicelles and siliceous oligomers hard to aggregate into a larger bulk. A feasible assembly process is speculated and shown in Scheme 2.

Scheme 2 Formation of sponge-like mesoporous silica.

Subsequently, basicity of materials was evaluated with Temperature Programmed Desorption of CO₂ (CO₂-TPD) (Fig. 4). The desorption peaks of CO₂ are provisionally divided into three components designated as α, β and γ at ca. 350, 470 and 720 °C, respectively, assigned to weak, medium, and strong basic sites. Obviously, all the curves of CaO-based HMS exhibit three kinds of different basic sites, suggesting the co-existence of weak, medium, and strong basic sites. It can be found that the effect of CaO concentration on basic intensity is weak. However, it is noticeable that the detected basic amounts in the sample show an obvious enhancing trend with increasing Ca concentration in the sample. When loadings of Ca in 5CaO–HMS reach a maximum, the γ basic sites observed at the high temperature of 718 °C are very abundant, indicating the presence of plentiful strong base sites in the template-induced CaO-based solid bases. It was reported that, in some traditional CaO-based mesoporous silica materials, the apparent weak basic sites with desorption temperature at ca. 150 °C are attributed to the SiO₂ matrix,³² which is in accord with the results from 5CaO–HMS(p) (Fig. 3S†), and the medium basic sites appearing at 220 °C correspond to surface CaSiO₃ species.³² However, in the desorption profiles of series xCaO–HMS, it can be observed that the present weak α basic sites at ca. 350 °C give an obviously higher basicity than those medium basic sites in traditional CaO-based materials. Herein, these weak α basic sites with an obvious enhanced basic intensity should be attributed to the CaO species directly connected to the siliceous pore wall, which is similar to CaSiO₃ species. Additionally, as expected, no CO₂ desorption peak attributed to the parent silica is detected, suggesting that the exposed silica surface is low due to the modification of dispersed CaO species, which favors a template-induced full coverage of CaO on the mesoporous wall. On the other hand, a relatively minor β peak at ca. 470 °C is observed and indexed to the medium basic sites of CaO species on the siliceous pore wall modified with calcium. According to some reports, generally, the highest CO₂ desorption temperature of pure CaO,³³ CaO/C,³³ CaO/ZrO₂ (ref. 34 and 35) and some other CaO-based silica materials^9,26 is only ca. 600 °C. However, herein, with respect to the series samples of xCaO–HMS, the desorption of CO₂ persists up to 800 °C and even higher temperatures, which is comparable to, not inferior to, some solid superbases, thereby demonstrating the presence of extraordinarily strong basicity in samples. It is concluded that, as compared to those reported Ca-based mesoporous materials prepared by doping and post-degradation methods, we have demonstrated that the template-induced Ca solid basic sites in this material are the strongest on the basis of their desorption temperature.

Fig. 4 Temperature-programmed desorption profiles of CO₂ over the series of xCaO–HMS mesoporous solid bases.

Fig. 5 displays the FT-IR spectra in the range from 1200 cm⁻¹ to 500 cm⁻¹ of the series samples of xCaO–HMS. The feature variations for the FT-IR spectra of xCaO–HMS can be obviously observed with increasing loadings of CaO on the surface. As shown, the feature absorption band attributed to the surface Si–OH groups at 962 cm⁻¹ exhibits a gradual drop in intensity with increasing incorporation of CaO species on the mesoporous surface. This is typical evidence for the consumption of silanol groups in the process of incorporation, owing to the strong interaction between the pore wall's silanol groups and the incorporated CaO. The almost disappeared silanol groups indicate the nearly complete coverage of CaO on the internal surface of mesoporous wall.^11,28 This coincides well with the CO₂-TPD results and is strong evidence for supporting the presence of thin-layer-like CaO on the surface of mesoporous wall.²⁸

Fig. 5 FT-IR spectra of samples of the series xCaO–HMS.

To accurately study the elemental chemical status and natural environment in samples and confirm the presence of Ca on the surface of DDA micelles, high-resolution XPS spectra for N 1s of u-5Ca–HMS and HMS–DDA, and Ca 2p of u-5Ca–HMS and 5CaO–HMS are recorded in Fig. 6a and b, respectively. As observed in Fig. 6a, the N 1s spectrum of the HMS–DDA shows a main peak at 399.60 eV and a shoulder at 401.3 eV, respectively, assigned to free amino groups of DDA³⁶ and its amino groups interacted with the mesoporous silicate surface by hydrogen-bonding.²⁸ Whereas the N 1s spectrum from u-5Ca–HMS gives two characteristic peaks at 399.71 and 401.68 eV, which departs obviously from that of hydrogen-bonding interacting with the siliceous wall. A slight shift (0.11 eV) for main component, an obvious shift (0.38 eV) for its shoulder to a higher binding energy and the enhanced intensity of its shoulder are discerned in the curve. This result obviously indicates a difference of the amino groups of DDA in terms of chemical state and coordination environment before and after introducing Ca²⁺ in the assembly process. This should be attributed to the coordination of Ca²⁺ on the N atoms of DDA micelles in which the electron-enriched nitrogen atoms donated electrons to Ca ions, and therefore the electron cloud densities from nitrogen atoms are reduced.³² This result confirms the presence and introducing of Ca²⁺ into the interface of the micelle and siliceous wall. On the other hand, the solid basic sites of Ca present in the materials are investigated by Ca 2p XPS spectra. The representative samples of u-5Ca–HMS and 5CaO–HMS were examined by high-resolution XPS and are displayed in Fig. 6b. Primarily, with respect to u-5Ca–HMS, binding energies from the Ca 2p transition splits at 346.95 and 350.5 eV for the 2p_3/2 and 2p_1/2 are observed, respectively. While in the profiles of 5CaO–HMS, binding energies from the Ca 2p transition splits at 347.6 and 351.4 eV corresponding to the 2p_3/2 and 2p_1/2 are observed, respectively, which is typically assigned to CaO.²⁶ Additionally, the obvious difference in binding energy from Ca atoms incorporating HMS before and after calcination suggests the changing chemical state of Ca atoms. This result manifests that the Ca species are not stably located in the framework of silica and surrounded by chemically-interacted Si–O species.²⁸ In fact, the Ca species should be introduced on the surface of the mesopore, supporting the fact of the template-induced dispersion.

Fig. 6 N 1s spectra (a) of u-5Ca–HMS (uncalcined 5CaO–HMS) and HMS–DDA (uncalcined HMS) and Ca 2p spectra (b) of 5CaO–HMS and u-5Ca–HMS.

3.3. Study of catalytic performance

Just as reported in the previous works, the transesterification process of PC and methanol (Scheme 3) could be effectively catalyzed with CaO-based bioactive catalysts at a lower temperature than other reported solid base catalysts need, and the reaction equilibrium can even be reached in 30 min.³⁷

Scheme 3 Transesterification reaction of PC and MeOH over CaO–HMS.

Herein, the catalytic performance of CaO-bearing HMS materials was evaluated with a transesterification reaction of PC and methanol, and the DMC and propylene glycol (PG) as the main products were directly synthesized in this reaction. Firstly, under a temperature of 50 °C, the effect of time on the transesterification reaction was investigated and the catalytic results are shown in Fig. 7a. It is noticeable that the template-induced 5CaO–HMS affords the faster reaction rate and higher conversion in the transesterification in comparison to the post-synthesized 5CaO–HMS(p). With regard to the superior catalytic activity of 5CaO–HMS, the large surface area and high exposure of the basic sites should be responsible for the results. On the other hand, to investigate the effect of temperature, the two catalysts were respectively used to catalyze the reaction under the different reaction temperatures ranging from 10 to 50 °C. The resulting catalytic data are displayed in Fig. 7b and c. The catalytic results reveal that the reaction over 5CaO–HMS can receive the optimized catalytic conversion under a lower reaction temperature (∼30 °C) in a shorter time. In contrast, the optimized catalytic conversion of post-synthesized 5CaO–HMS(p) requires a higher reaction temperature and longer reaction time. The short reaction time based on the 5CaO–HMS benefits from the large surface area and plentiful exposed active sites providing more diffused channels and more contact chances for the chemicals. While the lower reaction temperature may result from the enhanced basic intensity in the template-induced 5CaO–HMS, which is well supported by the reported work.³⁸

Fig. 7 Effect of reaction time based on 5CaO–HMS(p) and 5CaO–HMS (a): temperature: 50 °C. Effect of temperature on PC conversion and DMC selectivity with 5CaO–HMS, reaction time: 30 min (b), and 5CaO–HMS(p), reaction time: 60 min (c). All the reactions above were conducted under the conditions of atmospheric pressure (AP), MeOH/PC = 4 : 1, catalyst amount = 0.50 wt% of MeOH.

Subsequently, different calcium-containing HMS catalysts were tested under the optimized reaction conditions. The relationship of DMC yield versus different reaction times is shown in Fig. 8a. The detailed catalytic data using different catalysts are summarized in Table 1. As shown, the pure HMS gives negligible reaction activity even after a reaction for 30 min, which indicates the pure mesoporous silica is nearly inactive in the transesterification of PC. On the other hand, other calcium-containing solid basic catalysts exhibit certain catalytic activity, suggesting that calcium oxide should be the catalytic active center in this reaction. It is obvious that the catalytic activity of the series xCaO–HMS increases with increasing calcium loading in HMS, and the maximum yield (52.3%) of DMC is reached at the highest loading. This result demonstrates that the exposed calcium oxides on the mesoporous wall are increased with the Ca in the sample, coinciding with the ICP result in Table 1. It is also worth noting that the sequential increase of calcium loading in HMS leads to a slow increment of DMC yield at a high loading, indicating the nearly saturated status of surface calcium oxides on the mesoporous wall, and thereby confirming the aforementioned FT-IR analysis. In addition, the variation of DMC yield with different reaction times can be observed in Fig. 8a. The reaction equilibrium is almost reached in 25 min with respect to the series xCaO–HMS, indicating a quick reaction rate over the series xCaO–HMS catalysts. This is obviously attributed to the presence of a large surface and massive voids in the sample, which promote the rapid diffusion and reaction of chemicals. For comparison, the catalytic data of post-impregnated 5CaO–HMS(p) under 50 °C and 60 min is displayed in Table 1. However, even with higher reaction temperatures and longer reaction times, only a poor catalytic activity (30.4% of yield) is received with respect to the post-treatment one, which is obviously lower than that of the template-induced CaO-based material. Based on the high-angle XRD result of 5CaO–HMS(p), we can safely make a conclusion that a large amount of calcium oxide should be aggregated into large particles during the calcining process. Therefore, the exposed active sites on the support are limited even with a high loading. The result demonstrates that the high dispersion and exposure of active sites are extremely useful in heterogeneous catalytic reaction.

Fig. 8 The yield of DMC versus the reaction time over the different samples (a): temperature: 30 °C; catalytic performance of untreated CaO-based HMS under the respective optimized conditions and that of CO₂ treated CaO-based HMS under temperature: 50 °C, time: 60 min (b); relationship between catalytic activity and amount of γ basic sites in xCaO–HMS samples (c); stability test based on 5CaO–HMS and 5CaO–HMS(p) under their respective optimized conditions (d). All the reactions above were conducted under the conditions: atmospheric pressure (AP), MeOH/PC = 4 : 1, catalyst amount = 0.50 wt% of MeOH.

Table 1 Catalytic results for the synthesis of dimethyl carbonate with different catalysts^a

Samples	Ca/Si^b (%)	Conv.^c (%)	Sel.^d (%)	Yield (%)	Ref.
a Reaction conditions: 30 °C, 30 min, atmospheric pressure (AP), MeOH/PC = 4 : 1, catalyst amount = 0.50 wt% of MeOH.b The molar ratio of Ca/Si is determined from the result of ICP.c Conversion = mmol (PC) converted/mmol (initial PC) × 100%.d Selectivity of DMC = mmol (DMC)/mmol (PC converted) × 100%.e Reaction conditions: 50 °C, 60 min, MeOH/PC = 4 : 1, catalyst amount = 0.50 wt% of MeOH.
Pure HMS	—	4	—	—	This work
1CaO–HMS	1.84	15.9	92	14.6	This work
2CaO–HMS	3.56	27.7	93	25.8	This work
3CaO–HMS	5.23	39.8	91	36.2	This work
4CaO–HMS	7.11	51.6	90	46.4	This work
5CaO–HMS	8.87	56.8	92	52.3	This work
5CaO–HMS(p)^e	10.2	34.9	87	30.4	This work
CaO	—	53	87	46	33
CaO/C	—	—	—	42.9	33
CaO/ZrO2	—	45	—	—	34
Li2O/SiO2	—	32.4	—	—	39
K2O/SiO2	—	44.3	—	—	40

---

On the other hand, the reported reference catalysts were compared with the template-induced CaO-based HMS and their corresponding catalytic results for the synthesis of DMC with PC are listed in Table 1. Just as we observed, the CaO-based catalysts provided are based on different supports including support-free, carbon and ZrO₂; however, these reported CaO-based catalysts, even with a relatively high dispersity of active species, still exhibited relatively poor catalytic activity in comparison to the template-induced CaO-based HMS. Associated with the aforementioned analysis, the basic intensity of the template-induced CaO–HMS is obviously superior to that of CaO/C, even the pure CaO, so we suggest that the enhanced base sites may be responsible for the higher catalytic activity as compared to their counterparts. On the other hand, based on the same support, the Li₂O and K₂O modified mesoporous silica reported still do not show a higher yield for DMC, which may be due to their weaker basic sites compared to CaO–HMS, supporting the above analysis.

3.4. Catalytic study of special basic sites

In addition, to investigate the special effect of the γ basic sites in the synthesis of DMC, we conducted a designed experiment by a CO₂ absorption and desorption treatment at a specific temperature (550 °C) to cover γ basic sites and obtain CaO–HMS containing only α and β basic sites (Scheme 4). Wherein, the desorption temperature of CO₂ is dependent on the CO₂-TPD results and is lower than the desorption point of γ basic sites. The amount of γ basic sites is determined based on the titration method by subtracting the amount of α and β basic sites (from the treated sample under 550 °C) from the total basic amount (from the activated sample under 800 °C). Due to the coverage of enhanced γ basic sites, the CO₂ treated series xCaO–HMS was further used to catalyze the synthesis of DMC under the higher reaction temperature and longer reaction time (50 °C, 60 min). The received catalytic results are displayed in Fig. 8b. Just as we observed, the CO₂ treated series xCaO–HMS shows an obviously poor catalytic performance in comparison with the untreated series xCaO–HMS. Associating this with the aforementioned CO₂-TPD results, it can be known that the γ basic sites have been covered during the CO₂ absorption/desorption process. Therefore, herein the catalytic performance of the CO₂ treated series xCaO–HMS is merely attributed to the α and β basic sites left on the mesoporous wall. In addition, the relationship between the calculated DMC yields contributed from γ basic sites and the amount of γ basic sites (per g catalyst) is plotted and shown in Fig. 8c. This result demonstrates that the new special γ strong basic sites are very significant and contribute to the main catalytic activity in the synthesis of DMC. In addition, it is worth noting that, with the increment of calcium amounts in the samples, the CO₂ treated xCaO–HMS also shows increasing catalytic activity, which should be due to the increasing amount of the remaining α and β basic sites on the pore wall. On the other hand, it should be noted that CO₂ treated 5CaO–HMS(p) exhibits a higher catalytic activity as compared to the CO₂ treated 5CaO–HMS. This is because the strongest basic sites on the 5CaO–HMS(p) are much weaker than those of 5CaO–HMS (Fig. 3S†). Therefore, the low temperature desorption of CO₂ treated 5CaO–HMS(p) did not result in the whole coverage of the strongest basic sites of 5CaO–HMS(p). As a result, the more exposed basic sites give the higher catalytic activity with respect to CO₂ treated 5CaO–HMS(p).

Scheme 4 Coverage of γ strong basic sites and test of catalytic performance of remaining α and β basic sites.

3.5. Stability test

Finally, the catalyst of 5CaO–HMS as the optimized one is used for a stability test by recycled experiments, and the results are shown in Fig. 8d. Just as shown, the 5CaO–HMS exhibits a high stability and durable catalytic activity (≥47% of DMC yield) with six recycling experiments. However, as compared to the 5CaO–HMS, the post-synthesized sample 5CaO–HMS(p) fails to parallel the catalytic activity and stability in the synthesis of DMC. According to the aforementioned analysis, the obtained high stability may be attributed to the strong interaction produced between CaO and the mesoporous wall. Wherein, in this special templating method, the Ca atoms should be anchored on the silanol of the siliceous wall by certain physicochemical interaction.²⁸ Apart from that, the finally recycled 5CaO–HMS sample was examined by low-angle XRD, ICP, TEM and SEM techniques and the results are shown in Fig. 4S.† The results reveal that the structure of 5CaO–HMS has not been obviously destructed from the XRD, TEM and SEM results. However, the molar ratio of Ca/Si decreases slowly from 8.87% to 7.32% as detected by ICP results, which may be a reason (leaching of CaO) for the resulting slight drop in the catalytic activity.

3.6. Possible mechanism of reaction

A transesterification reaction between PC and methanol catalyzed with a base-catalyst is a typical nucleophilic substitution process.³⁷ In this transesterification process, methanol acts as nucleophilic regent, which would be activated by the CaO basic sites dispersed on the mesoporous wall. Subsequently, the activated methanol attacks the ester bond of PC and substitutes the opening sites of PC, in which the methanol would produce the absorption and dissociation interaction with the CaO species. The electronegative oxygen atom bonding to the Ca atom would capture the hydrogen ion of methanol; meanwhile, the positively-charged Ca atom would stabilize the alkoxy anion whose hydrogen ions are snatched away. Wherein, hydroxypropyl methyl carbonate (HPMC) is supposed to be the intermediates when the activated methanol attracts the PC. To further confirm the process of reaction, we calculated the charge distribution of methanol, PC and HPMC, and the distribution of Hirshfeld charge density and Coulomb potential patterns are shown in Fig. 5S.† The charge density of several special atoms are listed in Table 2. The hydroxyl oxygen atoms in methanol and HPMC carry a charge of −0.222 and −0.208, respectively. The charges of the carbonyl carbon atom in the HPMC and PC are 0.227 and 0.219, respectively. Thus, the reaction free energy of methanol and HPMC are significantly lower than the corresponding free energy of methanol and PC. Therefore, HPMC as an intermediate is regarded to be reasonable. On the basis of these analysis, we proposed a possible reaction mechanism of DMC synthesis and its depiction is displayed in Scheme 5.

Table 2 Hirshfeld charge distribution of MeOH, HPMC and PC molecule^a

Molecules	Charge of atoms
	Oa	Hb	Cc
a The charge distribution calculated from the Gaussian 09 package. The atoms Oa, Hb and Cc are illustrated as follows:.
MeOH	−0.222	0.151	—
HPMC	−0.208	0.148	0.227
PC	—	—	0.219

---

Scheme 5 Reaction process of PC to DMC and its possible reaction mechanism.

4. Conclusions

In summary, we successfully fabricated an unprecedented strong basicity of sponge-like mesoporous silica by the in situ incorporation of thin-layer CaO inside its mesopore based on a simple templating-induced method. An unexpected strong basicity is obtained and is obviously superior to that of traditional CaO-based silica material. Additionally, the addition of Ca²⁺ in the assembly process resulted in the formation of a new sponge-like microscale structure of silica. Based on its superior physicochemical properties, a high catalytic activity and stability for the synthesis of DMC was received. It was discovered that the enhancement of base strength could remarkably reduce the reaction temperature needed, and the special γ basic sites are responsible for the main catalytic activity.

Acknowledgements

The authors acknowledge project Funding by the financial support of the National Natural Science Foundations of China (21276125, 20876077 and 21476108), and the Priority Academic Program Development of Jiangsu Higher Education Institutions (PAPD).

Notes and references

1. M. Cargnello, V. V. T. Doan-Nguyen, T. R. Gordon, R. E. Diaz, E. A. Stach, R. J. Gorte, P. Fornasiero and C. B. Murray, Science, 2013, 341, 771–773.
2. F. A. Westerhaus, R. V. Jagadeesh, G. Wienhoefer, M.-M. Pohl, J. Radnik, A.-E. Surkus, J. Rabeah, K. Junge, H. Junge, M. Nielsen, A. Brueckner and M. Beller, Nat. Chem., 2013, 5, 537–543.
3. S. Pradhan, J. K. Bartley, D. Bethell, A. F. Carley, M. Conte, S. Golunski, M. P. House, R. L. Jenkins, R. Lloyd and G. J. Hutchings, Nat. Chem., 2012, 4, 134–139.
4. P. Puthiaraj and K. Pitchumani, Chem.–Eur. J., 2014, 20, 8761–8770.
5. G. Busca, Chem. Rev., 2010, 110, 2217–2249.
6. L. Zhu, F. Lu, X.-D. Liu, X.-Q. Liu and L.-B. Sun, Chem. Commun., 2015, 51, 10058–10061.
7. 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.
8. X.-Y. Liu, L.-B. Sun, F. Lu, X.-D. Liu and X.-Q. Liu, Chem. Commun., 2013, 49, 8087–8089.
9. L. B. Sun, J. H. Kou, Y. Chun, J. Yang, F. N. Gu, Y. Wang, J. H. Zhu and Z. G. Zou, Inorg. Chem., 2008, 47, 4199–4208.
10. Z. Y. Wu, Q. Jiang, Y. M. Wang, H. J. Wang, L. B. Sun, L. Y. Shi, J. H. Xu, Y. Wang, Y. Chun and J. H. Zhu, Chem. Mater., 2006, 18, 4600–4608.
11. F. Yang, S. Gao, C. Xiong, S. Long, X. Li, T. Xi and Y. Kong, RSC Adv., 2015, 5, 72099–72106.
12. F. Yang, S. Gao, C. Xiong, H. Wang, J. Chen and Y. Kong, Chin. J. Catal., 2015, 36, 1035–1041.
13. F. Yang, S. Long, S. Zhou, X. Li, X. Liu, S. Gao and Y. Kong, RSC Adv., 2016, 6, 30852–30861.
14. T. Yokoi, Y. Kubota and T. Tatsumi, Appl. Catal., A, 2012, 421, 14–37.
15. L.-B. Sun, X.-Q. Liu and H.-C. Zhou, Chem. Soc. Rev., 2015, 44, 5092–5147.
16. T.-T. Li, L.-B. Sun, L. Gong, X.-Y. Liu and X.-Q. Liu, J. Mol. Catal. A: Chem., 2012, 352, 38–44.
17. N. Kumar, E. Leino, P. Maki-Arvela, A. Aho, M. Kaldstrom, M. Tuominen, P. Laukkanen, K. Eranen, J.-P. Mikkola, T. Salmi and D. Y. Murzin, Microporous Mesoporous Mater., 2012, 152, 71–77.
18. L.-B. Sun, Y.-H. Sun, X.-D. Liu, L. Zhu and X.-Q. Liu, Curr. Org. Chem., 2014, 18, 1296–1304.
19. H. Jeon, D. J. Kim, S. J. Kim and J. H. Kim, Fuel Process. Technol., 2013, 116, 325–331.
20. G. Wang, L. Zhang, H. Dai, J. Deng, C. Liu, H. He and C. T. Au, Inorg. Chem., 2008, 47, 4015–4022.
21. X. G. Wang, K. S. K. Lin, J. C. C. Chan and S. F. Cheng, J. Phys. Chem. B, 2005, 109, 1763–1769.
22. D. Srinivas and P. Ratnasamy, Microporous Mesoporous Mater., 2007, 105, 170–180.
23. Z. Yan, G. T. Li, L. Mu and S. Y. Tao, J. Mater. Chem., 2006, 16, 1717–1725.
24. Y. L. Wei, Y. M. Wang, J. H. Zhu and Z. Y. Wu, Adv. Mater., 2003, 15, 1943–1945.
25. P. K. Raja, A. Chokkalingam, S. V. Priya, V. V. Balasubramanian, M. R. Benziger, S. S. Aldeyab, R. Jayavel, K. Ariga and A. Vinu, J. Nanosci. Nanotechnol., 2012, 12, 4613–4620.
26. M. C. G. Albuquerque, I. Jimenez-Urbistondo, J. Santamaria-Gonzalez, J. M. Merida-Robles, R. Moreno-Tost, E. Rodriguez-Castellon, A. Jimenez-Lopez, D. C. S. Azevedo, C. L. Cavalcante Jr and P. Maireles-Torres, Appl. Catal., A, 2008, 334, 35–43.
27. C.-H. Huang, K.-P. Chang, C.-T. Yu, P.-C. Chiang and C.-F. Wang, Chem. Eng. J., 2010, 161, 129–135.
28. F. Yang, S. Zhou, H. Wang, S. Long, X. Liu and Y. Kong, Dalton Trans., 2016, 45, 6371–6382.
29. F. Yang, S. Zhou, S. Gao, X. Liu, S. Long and Y. Kong, Microporous Mesoporous Mater., 2016 DOI:10.1016/j.micromeso.2016.03.007.
30. Z. H. Fu, J. H. Chen, D. L. Yin, D. H. Yin, L. X. Zhang and Y. Y. Zhang, Catal. Lett., 2000, 66, 105–108.
31. D. H. Yin, W. H. Li, W. S. Yang, H. W. Xiang, Y. H. Sun, B. Zhong and S. Y. Peng, Microporous Mesoporous Mater., 2001, 47, 15–24.
32. H. Sun, J. Han, Y. Ding, W. Li, J. Duan, P. Chen, H. Lou and X. Zheng, Appl. Catal., A, 2010, 390, 26–34.
33. T. Wei, M. H. Wang, W. Wei, Y. H. Sun and B. Zhong, Green Chem., 2003, 5, 343–346.
34. S. Liu, J. Ma, L. Guan, J. Li, W. Wei and Y. Sun, Microporous Mesoporous Mater., 2009, 117, 466–471.
35. H. Wang, M. H. Wang, W. Y. Zhang, N. Zhao, W. Wei and Y. H. Sun, Catal. Today, 2006, 115, 107–110.
36. H. Yu, Y. Zhu, H. Yang, K. Nakanishi, K. Kanamori and X. Guo, Dalton Trans., 2014, 43, 12648–12656.
37. T. Wei, M. H. Wang, W. Wei, Y. H. Sun and B. Zhong, Chin. J. Catal., 2003, 24, 52–56.
38. T. Wei, M. H. Wang, W. Wei, Y. H. Sun and B. Zhong, Fuel Process. Technol., 2003, 83, 175–182.
39. 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.
40. L. Zhu, F. Lu, X. D. Liu, X. Q. Liu and L. B. Sun, Chem. Commun., 2015, 51, 10058–10061.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra21236k

---