As the result of increasing interest in environmental protection, and because of associated governmental regulations, so-called green chemistry syntheses have received significant attention. Dimethyl carbonate (DMC) is a commonly used and environmentally benign building block in fine chemistry due to its useful chemical properties and low toxicity [1], and can be employed as a methylation or carbonylation reagent in place of the more hazardous substances phosgene and dimethyl sulfate [2]. DMC can also serve as an effective gasoline additive because of its high octane value [3], and is used as a monomer for the synthesis of polycarbonate resins [4].

Various methods has been developed for the synthesis of DMC, including the oxidative carbonylation of methanol [5, 6], the transesterification of methanol with cyclic carbonates [7-10], the alcoholysis of urea [11-13] and direct synthesis from carbon dioxide and methanol [14-16]. Among these, the transesterification of methanol with propylene carbonate (PC) is a promising route with the advantages of minimal potential for equipment corrosion, a high degree of safety and good yields [17, 18]. It has been determined that basic catalysts are effective for this reaction [19], and metal oxides are of particular interest due to their various advantages, such as high activity and ease of recycling during continuous processing [10, 20-23]. Calcium oxide exhibits good catalytic activity but poor stability [9, 24], and so the use of multi-metal composite oxides that facilitate the dispersion of CaO and increase the interactions of metals to prevent leaching has been proposed [9].

Layered double hydroxides (LDHs) have the general formula [M²⁺_1-xM³⁺_x(OH)₂]^x+[(A^n-)_x/n]^x-·mH₂O, where M²⁺ and M³⁺ represent divalent and trivalent metals and A^n- is an exchangeable interlayer anion [25]. Following the calcination of these materials, metal oxides with homogeneously dispersed M²⁺ and M³⁺ ions and adjustable acid-base properties are obtained [26]. Over the past several years, LDHs have been used in a wide array of practical applications in heterogeneous catalysis [26, 27], adsorption [28], pharmaceuticals [29], photochemistry and electrochemistry [25, 30]. Recently, studies have demonstrated that calcined Mg-Al, Ca-Al and Ca-Mg-Al LDHs are effective and stable transesterification catalysts [31−34]. Moreover, the introduction of basic rare earth elements (La, Ce or Y) into these materials can modify the basic sites and thus modulate the basicity of the catalyst [35].

In the present work, the basic metals Mg, La, Ce and Y were inserted into Ca-Al LDHs and the corresponding mixed oxides were obtained by calcination. The catalytic activities of these new materials were subsequently assessed during the transesterification of methanol with PC. The effects of these metals on the physicochemical properties of the as-prepared mixed oxides were also investigated using X-ray diffraction (XRD), thermogravimetry-differential thermogravimetry (TG-DTG), infrared (IR) spectroscopy, X-ray photoelectron spectroscopy (XPS), inductively coupled plasma-optical emission spectroscopy (ICP-OES), Brunauer-Emmett-Teller (BET) surface area analyses, and CO₂-temperature programmed desorption (TPD) and Hammett titration data. In this manner, a more comprehensive relationship between the catalytic activities of these materials and their physicochemical properties was developed.

Ca-M-Al (M = Mg, La, Ce or Y) LDHs were synthesized using a co-precipitation method. Typically, the appropriate amounts of metal chlorides (employing a (Ca+M)/Al molar ratio of 2 and a Ca/M molar ratio of 3) were dissolved in 100 mL distilled water to produce solution A. The required quantity of NaOH was also dissolved in 100 mL distilled water to give solution B. Subsequently, the two solutions were added dropwise by peristaltic pump to 250 mL of a water/ethanol solution having a water/ethanol volume ratio of 2:3, with vigorous stirring at 333 K under N₂ while maintaining the pH at 10, followed by stirring under the same conditions for 24 h. The precipitate was removed by filtration and washed with deionized water until the wash water showed a pH of 7, then dried under vacuum at 353 K and calcined at 1073 K under N₂ for 6 h. The precursors and calcined catalysts are designated herein as M-CAP and M-CA, respectively.

A pure Ca-Al LDH (with a Ca/Al molar ratio of 2) was synthesized using the same method described above, and the precursor and calcined catalyst are designated as CAP and CA, respectively. A mixture of CaO and Al₂O₃ was also prepared by mechanical mixing with a Ca/Al molar ratio of 2 and is designated as CA-mixing.

XRD data were acquired with a D8 Advance (Bruker, Germany) diffractometer, using Cu K_α (1.5406 Å) radiation at 40 kV and 50 mA. The scan rate was 3°/min in the 2θ range of 5°–80°. TG-DTG analyses were performed using a Rigaku TG 8120 instrument, heating each sample at 10 K/min from room temperature to 1173 K under N₂. Fourier transform IR (FT-IR) spectra were acquired with a Nicolet Nexus 470 FT-IR spectrophotometer over the range of 4000-400 cm^-1, summing 64 scans for each spectrum and with a resolution of 2 cm^-1. XPS data were acquired using an AXIS ULTRA DLD spectrometer with monochromatic Al K_α (1486.8 eV) radiation under ultrahigh vacuum. The binding energy (BE) values were calibrated internally using the C 1s peak with BE = 284.8 eV. The experimental error was within ±0.1 eV. Elemental chemical analysis was performed using ICP-OES with a Thermo iCAP 6300 instrument.

The BET specific surface areas and pore volumes of these materials were determined with a Micromeritics Tristar II 3020. CO₂-TPD data were acquired using a Builder PCA-1200 chemical adsorption instrument. Each sample was pretreated in a He flow at 1073 K for 1 h and then cooled to room temperature. CO₂ was adsorbed at room temperature for 30 min at 40 mL/min, after which a He flow was initiated to purge residual CO₂ for 30 min. The desorption was performed at a heating rate of 10 K/min from 323 to 1173 K, using a thermal conductivity detector and He as the carrier gas.

Hammett indicators were employed to determine the basicity and base strength of each catalyst. Pre-treated catalyst specimens were dispersed in cyclohexane and titrated with a solution of benzoic acid in cyclohexane in the presence of Hammett indicators to determine the total basicity. The base strength (H_) of each material was determined from the color change of various pH indicators after 12 h, using phenolphthalein (H_ = 9.3), 2, 4-dinitroaniline (H_ = 15.0), 4-nitroaniline (H_ = 18.4), 4-chloroaniline (H_ = 26.5) and diphenylmethane (H_ = 35.0).

All catalysts were activated at 1073 K under N₂ for 1 h prior to catalytic trials in a glass batch reactor equipped with a spiral condenser at 333 K. Following each reaction, the product was centrifuged, filtered and analyzed with a gas chromatograph equipped with a flame ionization detector. A stoichiometric excess of methanol was used during these reactions, and the PC conversion, DMC selectivity and DMC yield were calculated by the following equations.

(1)

(2)

(3)

Here, m_PC1 is the mass of PC in the feed, m_PC2 is the mass of PC in the product, m_DMC is the mass of DMC generated, M_PC is the molar mass of PC, and M_DMC is the molar mass of DMC.

The XRD patterns of the CAP and M-CAP samples are shown in Fig. 1(a). It is evident that all samples had a typical layer structure and generated three sharp, symmetric characteristic peaks that can be assigned to the reflections of the (002), (004) and (020) planes. The peak intensities of the samples are different due to variations in the valances and dimensions of the ions. The Mg-CAP shows excellent crystallinity compared with the CAP sample, suggesting that Mg²⁺ favors the formation of the LDH structure. However, the peak intensities of the other M-CAP samples are lower than those of the CAP, demonstrating comparatively poor crystallinity.

Following calcination, all samples were in the form of composite oxides (Fig. 1(b)). A mayenite (Ca₁₂Al₁₄O₃₃) phase is observed in the pattern of each sample, indicating an interaction between Ca and Al. In the case of the Mg-CA, the intensity of the CaO peak is much higher than those of the other metal-modified samples. Moreover, the chemical interaction between Mg and Al evidently generated a new MgAl₂O₄ phase. It is also apparent that La interacted with Ca and Al to form a CaLaAlO₄ phase, while Y only interacted with Al to form an Al₅Y₃O₁₂ phase. Interestingly, the Ce-CA sample demonstrates a CeO₂ phase with a weak Ca₁₂Al₁₄O₃₃ peak, possibly because the introduction of Ce leads to a disordered structure.

The TGA-DTG data acquired from the CAP and M-CAP samples are depicted in Fig. 2, and indicate the typical thermal decomposition stages of LDH structures [36]. In the case of the CAP sample, three major mass losses are observed [36]. The first occurs over the range of 300-426 K and corresponds to the removal of physically absorbed and interlayer water. The second mass loss in the range of 548-583 K can be attributed to the de-hydroxylation of the inorganic layers, leading to collapse of the layered structure, while the final mass loss at 871-983 K can be ascribed to the decomposition of the subject-object structure. However, it is clear that the M-CAP samples behave differently. The decomposition of these materials occurs at lower temperatures, especially that of the Ce-CAP, suggesting that the insertion of metal ions may distort the main structure and decrease the thermal stability of Ca-Al LDHs.

FT-IR spectroscopy was employed to study the structural features of CAP and CA samples (Fig. 3). The O-H stretching vibration of free water at 3644 cm^-1 can be observed [12] in these spectra, and the stretching mode of hydroxyl groups in layer and interlayer water molecules appears as a broad band at approximately 3455 cm^-1 [37]. The band at 1621 cm^-1 is due to the bending vibration of physically absorbed water [38]. The band at 1400 cm^-1 demonstrates the presence of mono and bidentate carbonates, even though carbonates were not detected by XRD, confirming that the catalyst surfaces were subjected to a high degree of carbonation [39, 40]. In addition, bands attributed to metal-OH bonds are also evident at 782, 526 and 427 cm^-1.

After calcination at 1073 K, all samples show considerable changes. A broad but less intense band appears at 3460 cm^-1 and can be assigned to the stretching mode of hydroxyl groups. Weak bands originating from surface carbonate or hydrogen carbonate species are also present as a result of the adsorption of atmospheric CO₂ on the external surfaces of the catalysts during handling. The band at 1410 cm^-1 can be ascribed to the O−C−O stretching vibrations of adsorbed carbonate anions on the surface basic sites of the catalysts. The intensity of this peak in the Mg-CA and La-CA spectra is higher than in the case of the pure CA sample, suggesting a greater carbonate content on the surfaces of the former materials than on the pure CA, because of higher surface basicity. Therefore, introducing the appropriate amounts of Mg or La may increase the surface basicity of these materials [41]. In the low wavenumber region (400-1000 cm^-1), the bands related to metal−OH bonds shift to higher wavenumbers. It is also apparent that the spectra are dominated by a strong band at 852 cm^-1 (attributed to metal−oxygen bonds) that is absent in the patterns obtained from the CAP and M-CAP samples, suggesting that metal oxides are formed during calcination. Variations in the shapes and intensities of these bands between specimens may originate from the formation of different oxide species.

The surface properties and compositions of the CA and M-CA catalysts were assessed by XPS, and the data in Table 1 show that the Ca and Al BE values are very similar for all samples. Fig. 4 provides the XPS spectra of the CA and M-CA catalysts. The Mg-CA Mg 2p spectrum contains a single peak at 50.0 eV that is consistent with the presence of Mg²⁺ [42]. In addition, the La 3d_5/2 core-level peak is located at a BE of 834.8 eV, a value that is higher than those reported for La₂O₃ (834.3 eV) [43] or LaAlO₃ (833.8 eV) [44]. This peak can therefore be assigned to well-dispersed La species [45]. In the Ce 3d spectrum, the three main Ce 3d_5/2 features at 882.1, 888.1 and 897.8 eV and the three main Ce 3d_3/2 features at 900.3, 907.0 and 916.2 eV can be ascribed to Ce⁴⁺ in CeO₂ [46], which is consistent with the XRD results. Two other peaks, at 883.6 and 901.3 eV, are also evident and are attributed to Ce³⁺. In the case of the Y-CA catalyst, four peaks are present at 156.4, 157.7, 158.4 and 160.1 eV, corresponding to Y 3d_5/2 and Y 3d_3/2 electrons [47, 48]. Compared with the peaks generated by pure Y₂O₃ at approximately 156.8 and 158.9 eV [49], the higher Y 3d BE values (157.7 and 160.1 eV) suggest that the Y atoms have combined with atmospheric moisture to produce Y−OH or Y−OC sites [48]. The O 1s spectra of the CA and M-CA catalysts contain three intense peaks at 530, 531 and 532 eV, ascribed to lattice O^2-, surface hydroxyl groups and carbonate groups, respectively [50, 51]. It can be seen that the O 1s BEs of the catalysts modified with Mg, La or Ce are shifted to lower values (Table 1), and so the electron pair donating ability of the surface oxygen atoms may be enhanced [52, 53].

The textural parameters and metal compositions of the catalysts are summarized in Table 2. Compared with the CA catalyst, the specific surface areas and pore volumes of the catalysts were increased with the introduction of Mg, La or Ce. In addition, the surface concentrations of the metals and of Ca are noticeably higher, suggesting that the catalyst surfaces were enriched in these elements. Notably, the Mg-CA and La-CA catalysts exhibit lower Al levels compared with the other catalysts.

The basicity of each catalyst was determined using CO₂-TPD and Hammett indicators. The CO₂-TPD profiles of the CA and M-CA catalysts are shown in Fig. 5. For all catalysts, the broad peak ranging from 300 to 700 K can be divided into two peaks (peak α and peak β), attributed to weak and moderate basic sites that primarily result from basic OH^- groups and M^x+-O^2- pairs. Another broad peak (peak γ) between 700 and 900 K can be ascribed to strongly basic sites associated with unsaturated O^2- anions. The CA catalyst shows a symmetric desorption peak at 800 K, whereas the M-CA generates more complex data in which more than two Gaussian peaks are present, suggesting that the addition of metal ions has a significant effect on the strongly basic surface sites. Prior studies have demonstrated that the peaks above 900 K can be assigned to super strong basic sites [54]. The CO₂-TPD profiles confirm that super strong basic site (peak δ) were present on the Mg-CA and La-CA catalysts. Therefore, the introduction of Mg and La evidently generates a certain number of super strong basic sites, possibly due to the rearrangement of isolated O^2- ions on the catalyst surfaces. Wu et al. [55] reported that the M^x+-O^2- pairs on the catalyst surface could partially split to generate unsaturated O^2- ions. The Mg-CA catalyst was found to have the highest (Mg+Ca):Al atomic ratio among the catalysts (Table 2), which may be responsible for the presence of more unsaturated O^2- ions, leading to the highest concentration of super strong basic sites.

The base strength (H_) of each catalyst was subsequently assessed using Hammett indicators. The basicity distributions of the catalysts are summarized in Table 3. According to the results of titrations, three basic sites were identified: weak (9.3-15.0), medium (15.0-18.4) and strong (18.4-26.5). Noticeably, values in the super strong basic strength range (26.5-35.0) were found with the introduction of Mg and La, which is consistent with the CO₂-TPD results.

The total basic amounts determined from the Hammett indicators were relatively high compared with those obtained from CO₂-TPD, because CO₂-TPD primarily responds to surface basic sites, while titrations with Hammett indicators can detect both exposed and internal basic sites. The total amounts of basic sites determined by both methods were in the same order of Y-CA < CA < Ce-CA < La-CA < Mg-CA.

The effects of the methanol/PC molar ratio and catalyst mass on DMC yields were investigated (Fig. 6). The data demonstrate that the yield increases along with the methanol/PC molar ratio within a certain range and then falls somewhat at a ratio of 15 because the PC concentration becomes too low for the transesterification to proceed efficiently [56]. In contrast, the yield only exhibits a slight increased up to a catalyst mass of 2 wt% with minimal changes above that level, suggesting that reaction equilibrium has been reached [57]. Therefore, the optimum reaction conditions consist of a methanol/PC molar ratio of 12 and a catalyst mass of 2 wt%. The catalytic activity increases in the order of Y-CA < CA < Ce-CA < La-CA < Mg-CA, and the highest PC conversion of 55.3% and DMC selectivity of 96.3% were obtained over the Mg-CA catalyst (Table 4).

During the reaction, the role of the basic sites is to activate CH₃OH to form CH₃O^-, which then attacks PC to produce DMC. Many studies [10, 58, 59] have reported that catalysts having the highest basic amounts will exhibit the best catalytic performance, and that high basic strength is also responsible for promoting the reaction [59]. In this study, the introduction of Mg, La or Ce increased the total surface basic amount of the catalysts. This is also reflected by increases in the DMC yields. The activity of the catalysts increases in the order of Y-CA < CA < Ce-CA < La-CA < Mg-CA, which is in good agreement with the basicity of the catalysts as calculated using the CO₂-TPD results. Thus, the relationship between the surface basicity and DMC yield was investigated. Fig. 7 demonstrates that the total surface basic amount and density of the as-prepared catalyst both have a linear relationship with the DMC yield.

Fig. 8 presents the results of recyclability trials using the catalysts. Although the initial activity of the CA-mixing catalyst is the highest, the activity deceases drastically after the tenth recycling. This effect may have resulted from the weak interaction between Ca and Al in the CA-mixing sample. In the case of the CA catalyst, the DMC yield decreases by 12.6% after 10 recycling runs, suggesting that the recyclability of this materials has been improved owing to the increased interaction between Ca and Al. Following modification with metal ions, the M-CA catalysts exhibit a longer recycle life compared with the CA-mixing and CA catalysts, such that the catalytic performance of the former is higher than those of the latter after five recycling runs. Finally, the recyclability of the catalysts after 10 recycling runs increases in the order of CA-mixing < CA < Ce-CA < Y-CA < La-CA < Mg-CA. This excellent recyclability may be ascribed to the intermetallic interactions induced by the addition of metals, which stabilize the active species of the catalysts.

The XRD patterns of the fresh and ten times recycled Mg-CA catalyst are provided in Fig. 9. It is evident that there is no difference in the patterns after ten recycling trials, suggesting significant stability. The chemical compositions of the fresh and recycled samples are also summarized in Table 5. The decrease in the catalytic performance after 10 recycling runs can be ascribed to a slight loss of active Ca and Mg ions.

LDHs containing various alkaline earth or rare earth metals Ca-M-Al (M = Mg, La, Ce, Y) were prepared and applied to a transesterification reaction after calcination. XRD analysis results indicate that Ca-M-Al LDHs were successfully synthesized and subsequently decomposed to give the corresponding metal oxides during calcination. The surfaces of these catalysts were rich in Mg, La, Ce or Y as well as in Ca, as demonstrated by XPS data. The results from basicity analyses by CO₂-TPD and Hammett titrations show that the basicity of the catalysts is increased following modification with Mg, La or Ce. The total surface basic amount and density both exhibited a linear correlation with the DMC yield. Furthermore, the recyclability of the M-CA catalysts was improved, such that a Ca-Mg-Al specimen maintained the highest activity after 10 recycling trials, with no appreciable variations in its XRD pattern with repeated use. The work reported herein is expected to lead to the development of high-efficiency solid base catalysts for the transesterification of PC with methanol. Such catalysts offer a simple, ecologically safe, cost effective and energy efficient means of synthesizing DMC while also allowing facile recycling.