Ethanol (EtOH) is regarded as a renewable platform chemical for the synthesis of versatile high value-added products [1-4], due to its low cost, large-scale industrial manufacture from biomass fermentation [5, 6]. Diethyl carbonate (DEC) is an important derivative of ethanol that is attracting increasing interest as a synthetic intermediate and green solvent [7], with the merits of low toxicity, negligible pollution, and low bio-accumulation. Moreover, it is an ideal substitute for the traditional fuel additive methyl tert-butyl ether (MTBE), due to its higher oxygen content and superior gasoline/water distribution coefficient [8]. To date, several approaches have been developed for the synthesis of DEC from ethanol: the reaction of ethanol with (a) phosgene [9], (b) urea [10-13], (c) carbon monoxide [14, 15], and (d) a dialkyl carbonate [16-18]. Among these methods, ethanolysis of urea has attracted a great deal of attention due to the advantages of cost effectiveness and facile manipulation and product separation processes [19, 20]. This reaction is comprised of two consecutive steps: the formation of intermediate ethyl carbamate (EC) from ethanol and urea, followed by reaction of EC and ethanol to produce DEC. Traditionally, organometallic compounds, metal salts, and metal oxides have been employed for ethanolysis of urea and its analogs, among which zinc oxide (ZnO) has shown a high DEC yield. It has been determined that the acid-base properties dominate the catalytic performance of ZnO [10, 21, 22], but the role of the acidic and basic sites in the ethanolysis of urea remains ambiguous. Moreover, the ZnO catalyst suffers from dissolution and transformation to the homogeneous complex Zn(NCO)₂(NH₃)₂ during the reaction process [23]. This stimulated us to further tune the acid-base properties of ZnO-based catalysts to achieve significantly improved catalytic performance and stability.

Layered double hydroxides (LDHs) (also known as hydrotalcites), with the general formula [M_1-x²⁺M³⁺_x(OH)₂](A^n-)_x/n·mH₂O, are a type of anionic clay with positively charged, brucite-like host layers, and exchangeable interlayer anions [24, 25]. Recently, LDH materials, as solid acid-base catalysts and catalyst supports, have received a great deal of interest in the field of heterogeneous catalysis, due to their versatile, tunable chemical compositions, structures, and architectures [26, 27]. In particular, by virtue of the so-called structural topotactic transformation process of LDH precursors, corresponding mixed metal oxide (MMO) catalysts can be obtained after calcination treatment [28], which possess tunable acid-base sites which depend on the chemical composition of, and preparation parameters from LDHs. This unique property inspired us to design and prepare MMO catalysts with tunable acid-base sites, for the purpose of attaining enhanced catalytic behavior toward the synthesis of DEC from ethanol and urea.

In this study, Zn₂Zr_xAl-hydrotalcite precursors (denoted as Zn₂Zr_xAl-LDH) were synthesized using a co-precipitation method. Subsequently, a series of Zn-Zr-Al mixed metal oxide (Zn₂Zr_xAl-MMO) catalysts were synthesized by phase topotactic transformation from LDH precursors and used in ethanol conversion to produce DEC. CO₂ temperature programmed desorption (CO₂-TPD) and NH₃-TPD studies indicated that the Zr content significantly influences the surface acidity and basicity of the Zn₂Zr_xAl-MMO catalysts, and the Zn₂Zr_0.1Al-MMO sample showed the maximal catalytic performance (DEC yield: 42.1%), which is the highest in comparison with those of previously reported metal oxide catalysts (Table S1). In addition, the Zn₂Zr_0.1Al-MMO catalyst exhibits good stability and reusability, with a loss in DEC yield of below 5% for up to five cycles. A study of structure-property correlations was performed: in situ FTIR results confirm that the weak acid site of ZrO₂ is responsible for the activation adsorption of urea and the intermediate product, while the medium-strong basic site of ZnO accelerates ethanol activation.

Analytical-grade chemicals, including ZrO(NO₃)₂·xH₂O, Zn(NO₃)₂·6H₂O, Al(NO₃)₃·9H₂O, Na₂CO₃, NaOH, ethanol, cyclohexanol, and urea were obtained from Sigma-Aldrich and all reagents were used without further purification.

Zn₂Zr_xAl-LDH precursors with Zn : Zr : Al molar ratios of 2 : 0 : 1, 2 : 0.05 : 1, 2 : 0.1 : 1, and 2 : 0.2 : 1 were prepared using a co-precipitation method, as follows. Typically, Solution A was obtained by dissolving ZrO(NO₃)₂·xH₂O, Zn(NO₃)₂·6H₂O, and Al(NO₃)₃·9H₂O, with varying Zn : Zr : Al molar ratios ([Zn ²⁺]+[Al³⁺]+[Zr⁴⁺] = 0.05 mol L^-1), in 100 mL of deionized water. Solution B was a mixture of NaOH (2 mol L^-1) and Na₂CO₃ (0.3 mol L^-1) in deionized water (100 mL). Solution B was added dropwise to Solution A with vigorous agitation until pH = 10 was obtained. The resulting slurry was heated at 110 ℃ for 12 h. The resulting precipitate (Zn₂Zr_xAl-LDH precursors) was separated, washed thoroughly with deionized water, and dried at 60 ℃ for 24 h. Finally, the Zn₂Zr_xAl-MMO samples were obtained via calcination of Zn₂Zr_xAl-LDH precursors at 500 ℃ for 4 h in air (heating rate 5 ℃ min^-1).

Powder X-ray diffraction (XRD) patterns were performed on a Rigaku XRD-6000 instrument (Cu K_α radiation: λ = 0.15418 nm; 40 kV, 30 mA). The UV-vis diffuse reflectance spectra (UV-vis DRS) were recorded on a Hitachi UV-vis spectrophotometer (U-3900H). The morphological study was carried out on a Zeiss Supra 55 scanning electron microscope (SEM) (accelerating voltage 20 kV). High resolution transmission electron microscopy (HRTEM) observation was carried out on a JEOL JEM-3010 at an accelerating voltage of 200 kV. Low-temperature N₂ adsorption-desorption isotherms were recorded on a Quantachrome Autosorb-1C-VP analyzer. The sample was degassed at 200 ℃ for 12 h; the multipoint Brunauer-Emmett-Teller (BET) method was used to calculate the total surface area.

The acidity and basicitiy of the catalysts were determined by using temperature-programmed desorption (TPD) of chemisorbed NH₃ and CO₂, carried out on a Micromeritics AutoChem Ⅱ 2920 apparatus equipped with a thermal conductivity detector (TCD). For the CO₂ or NH₃-TPD, 100 mg of sample was placed in a quartz tube reactor and pretreated in a helium flow (50 cm³ min^-1) at a heating rate of 10 ℃ min^-1 to 500 ℃, and maintained at 500 ℃ for 1 h, followed by cooling to 100 ℃. Afterwards, a flow of CO₂ or NH₃ was introduced into the sample for 0.5 h to obtain a saturation adsorption. Then, the catalyst was degassed in a helium flow for 1.5 h to remove physisorbed CO₂ or NH₃, followed by desorption measurements of NH₃ or CO₂ by heating from 100 to 550 ℃ at a rate of 10 ℃ min^-1.

In situ Fourier transform infrared (FTIR) spectra of urea, ethanol, and intermediate (EC) adsorbed onto catalyst samples were carried out using a Thermo Nicolet 380 spectrometer. Two types of measurements were performed, as follows. For urea or EC adsorption, the catalyst (20 mg) and urea (or EC 2 mg) were mixed thoroughly and pressed into a wafer. In the case of ethanol adsorption, the catalyst sample (20 mg) was pressed to obtain a wafer, which was then installed in the IR cell. Afterwards, ethanol was introduced into the cell for signal recording in the temperature range 30-200 ℃.

Catalytic reactions were performed using a 50-mL stainless-steel autoclave equipped with a magnetic stirrer. Typically, 0.08 g of catalyst, 0.8 g of urea, and 9.2 g of ethanol were placed in the reactor, which was purged three times with N₂, and then sealed. The reaction was triggered by heating the autoclave to a targeted temperature. The resulting product mixture was identified using off-line gas chromatography (Shimadzu GC-2014C, flame ionization detector). All detected products were quantitatively analyzed by the internal standard method using cyclohexanol as the internal standard.

A recyclability study of the catalyst was performed. The used catalyst was filtered from the mixed products, washed several times with ethanol and then dried. The obtained catalyst was reused for the next reaction cycle.

The yield of DEC (or EC) was determined by the following formula:

(1)

where M stands for the cumulative mole number for DEC (or EC).

The catalyst precursors, Zn₂Zr_xAl-LDH with varying Zr contents, were prepared via a facile co-precipitation method. Their XRD patterns (Fig. 1 (A)) display a series of reflections at 2θ values of 12°, 24°, 34.9°, 39°, 61°, and 62°, which were indexed to the (003), (006), (009), (015), (110), and (113) reflection planes of an LDH phase. No other crystalline phase was detected, implying high purity of the LDH precursors. However, as the Zr content increased, the crystallinity of Zn₂Zr_xAl-LDH decreased, indicated by weakened reflection intensity, which is due to the distortion of the LDH host caused by the introduction of the Zr⁴⁺ cation into the brucite-like layers. Fig. 1 (B) illustrates XRD patterns of the as-obtained Zn₂Zr_xAl-MMO samples after calcination treatment of the LDH precursors at 500 ℃. In all cases, LDH samples undergo a structural topotactic transformation process, resulting in the disappearance of the characteristic reflections of the LDH; while a series of reflections at 2θ values of 31.6°, 34.2°, 36°, 47.3°, 56.5°, 62.6°, and 67.8° are observed, corresponding to the ZnO phase (PDF#36-1451). No zirconium oxide or alumina oxide phase was detected, suggesting an amorphous phase.

UV-vis diffuse reflectance spectroscopy (Fig. 2) was employed to study the structure of the Zn₂Zr_xAl-MMO samples. All the samples showed an absorption peak at 375 nm, which is ascribed to the ZnO phase resulting from the transition of O (2p)-Zn (4s-p) [29]. A weak band at ~270 nm was observed in the three samples incorporating Zr and is attributed to ZrO₂ nanoparticles [30]. It should be noted that with the increase in Zr content, the band intensity at ~270 nm gradually increased, indicating the presence of ZrO₂ in the Zn₂Zr_xAl-MMO (x = 0.05, 0.1, 0.2) samples.

SEM images of the Zn₂Zr_xAl-LDH precursors show a uniform nanoplate-like shape with a lateral diameter distribution of 180-200 nm; the Zn₂Zr_xAl-MMO samples inherit the original plate-like morphology of LDHs without an obvious change in particle size (Fig. S1). As a typical example, TEM images of Zn₂Zr_0.1Al-LDH and Zn₂Zr_0.1Al-MMO display sheet-like shapes with particle sizes of 180-200 nm (Fig. 3 (A) and 3 (B)). A lattice fringe with a spacing of 0.247 nm was observed for the Zn₂Zr_0.1Al-MMO sample, which corresponds to the (101) reflection plane of the ZnO phase (Fig. 3 (C)). No Al₂O₃ reflections were observed, indicating an amorphous phase. Moreover, we did not observe the lattice fringe of ZrO₂, possibly due to its low content and poor crystallinity. The nitrogen adsorption-desorption isotherms (Table 1) of the Zn₂Zr_xAl-MMO samples indicated comparable specific surface areas, average pore sizes, and total pore volumes.

The conversion of ethanol with urea to produce DEC occurs in two consecutive steps: formation of the intermediate product EC, followed by reaction of EC with ethanol to produce DEC [31]. We studied the influence of the Zn/Al ratio on catalytic performance, and the results showed that the optimal catalytic behavior was obtained with Zn/Al = 2. Catalytic evaluations of the Zn₂Zr_xAl-MMO samples for ethanol conversion to DEC were performed (Fig. 4 (A), Tables 1 and S3). It was observed that the Zn₂Al-MMO catalyst showed the lowest yield of DEC (28.3%) among the catalysts. As the Zr content increased, the DEC yield first increased and then declined, and the Zn₂Zr_0.1Al-MMO catalyst exhibited the maximum yield of DEC (42.1%). This is, to the best of our knowledge, the highest value reported for metal oxide catalysts (Table S1) [10, 13, 21, 31, 32]. Furthermore, the recyclability and stability of the Zn₂Zr_0.1Al-MMO catalyst were also studied (Fig. 4 (B)). The DEC yield showed a slight decrease of ~2% after five reaction cycles. In addition, the XRD pattern of the used catalyst (Fig. 4 (C)) displayed no significant variation from that of the fresh catalyst. Nitrogen adsorption-desorption isotherms (Table S2) indicate an ~16% decrease in specific surface area after five reaction cycles, demonstrating satisfactory reusability.

It has been reported that the acid-base properties of the catalyst are crucial for alcohol conversion [34, 36]. Herein, CO₂-TPD and NH₃-TPD measurements were performed (Fig. 5 and Table 2) in order to determine the concentration of the acidic and basic sites and the strength of the acid-base properties of the Zn₂Zr_xAl-MMO catalysts. In the CO₂-TPD curves, all four samples display a broad CO₂ desorption peak between 100 and 400 ℃, which were deconvoluted into three contributions with maximum temperatures (T_M) in the ranges of 120-140 ℃ (weak base site), 180-200 ℃ (medium-strong base site) and 260-280 ℃ (strong base site) using a Gaussian peak fitting method [35-37]. The basic sites originate from OH groups (weak), M⁺-O^2- pairs (medium), and isolated O^2- ions (strong), respectively. In the case of the NH₃-TPD measurements, the NH₃ desorption peak in the range of 100-450 ℃ was fitted to three peaks with T_M in the region of 140-160 ℃ (weak acid site), 220-240 ℃ (medium-strong acid site), and 320-360 ℃ (strong acid site), respectively. The acid sites were derived from the contributions of a variety of metal oxides, such as Al₂O₃ and ZnO. It should be noted that both the total basic site concentration (B_TOTAL) and total acid site concentration (A_TOTAL) increased gradually from Zn₂Al-MMO to Zn₂Zr_0.1Al-MMO, and then decreased in Zn₂Zr_0.2Al-MMO, indicating the Zr content plays a key role in tuning the acid-base properties of the Zn₂Zr_xAl-MMO catalysts. With an increase in Zr content, the concentration of the medium-strong base sites (B_M) and the weak acid sites (A_W) increased significantly and reached a maximum in the Zn₂Zr_0.1Al-MMO catalyst. Moreover, we studied the DEC yield as a function of unique structure concentration, and the results showed a positive correlation between DEC yield and B_M and A_W (Fig. 5); however, this is not observed for other types of base or acid sites. These results imply that both medium-strong base and weak acid sites contribute to catalytic performance in the synthesis of DEC from ethanol and urea. The used Zn₂Zr_0.1Al-MMO was characterized by CO₂-TPD and NH₃-TPD, and the ratio of desorption peaks did not show an obvious change (Fig. S2).

We prepared a control sample (Zn₂Al-MMO) for a comparison study. It was found that the Zn₂Al-MMO sample produced the highest yield of EC (54.7%) and the lowest yield of DEC (28.3%) among all the catalysts (Table 1), indicating that the intermediate EC is not effectively activated and converted into DEC in the absence of ZrO₂. Moreover, both ZrO₂ and Al₂O₃ show high yields of EC (68.0% and 81.2%) but low yields of DEC (8.7% and 5.3%), as shown in Table S1. This indicates that the ethanol molecule does not undergo effective activation so as to react with EC in the absence of ZnO. We identified weak acid sites on the surface of ZrO₂ and medium-strong basic sites on ZnO based on the structure-property correlations in this work, as well as in previous studies. Based on the in situ FTIR absorption spectra of control samples, it is proposed that urea and intermediate EC are activated by weak acid sites (ZrO₂); whereas ethanol is activated by medium-strong basic sites (Zn²⁺-O^2-).

Furthermore, in situ FTIR absorption spectroscopy was employed to study the role of the acid sites of Zn₂Zr_0.1Al-MMO (Fig. 6 (A)), in which urea and intermediate EC were used as basic species to probe the acid sites. For the adsorption of urea on the Zn₂Zr_0.1Al-MMO sample, at room temperature, multiple bands were observed at 3456, 3355, 1683, and 1152 cm^-1, which were attributed to N-H, C=O, and C-N stretching vibrations, respectively [38, 39]. As the temperature increased to 130 ℃, a new band appeared at 2207 cm^-1, corresponding to ν_N=C=O bonding to the surface of the metal oxide (O=C=N-M, M stands for metal cation) [40], which originated from the activation of urea on the acid sites of the Zn₂Zr_0.1Al-MMO catalyst at high temperature. Additionally, the intensity of the band at 2207 cm^-1 increased significantly as the temperature increased from 130 to 200 ℃, demonstrating that a higher temperature is favorable for the activation adsorption of urea. In the in situ FTIR absorption of intermediate EC on the catalyst surface (Fig. 6 (B)), a similar phenomenon was observed; the band at 2207 cm^-1, due to O=C=N-M, was detected at 130 ℃ and then increased gradually between 130 and 200 ℃, as a result of the adsorption of EC. The results confirm that both urea and EC undergo activation adsorption on the acid sites of the Zn₂Zr_0.1Al-MMO catalyst.

In situ FTIR absorption spectra of ethanol on the surface of Zn₂Zr_0.1Al-MMO were carried out (Fig. 7), in which ethanol was used as an acidic species to detect basic sites. In the temperature range of 100-200 ℃, five bands, attributed to ethoxide species, were observed at 2973, 2925, 2900, 1078, and 1052 cm^-1, which are assigned as follows: ν_as(CH₃) stretching (2973 cm^-1); ν_as(CH₂) stretching (2925 cm^-1); ν_s(CH₃) stretching (2900 cm^-1); monodentate ethoxide stretching ν(CO_mono) (1078 cm^-1), and bidentate ethoxide stretching ν(CO_bi) (1052 cm^-1) [41, 42]. The ethoxide species originated from deprotonated ethanol (dissociation adsorption) on the basic sites of the Zn₂Zr_0.1Al-MMO catalyst. Moreover, the band intensities of ethoxide species became stronger with an increase in temperature, suggesting an enhanced activation adsorption of the ethanol molecule on the catalyst surface.

In summary, ZnZrAl-MMO catalysts were synthesized by the structural topotactic transformation process of LDH precursors. The Zn₂Zr_0.1Al-MMO catalyst exhibits the optimal catalytic performance (DEC yield 42.1%) for ethanol conversion to synthesize DEC. The CO₂-TPD and NH₃-TPD studies indicate that weak acid and medium-strong base sites account for the enhancement of the DEC yield. According to in situ FTIR investigations, the acid sites of the Zn₂Zr_0.1Al-MMO catalyst activate urea and the intermediate EC to produce O=C=N-M; while the base sites promote activation adsorption of ethanol, with the formation of ethoxide species. This study demonstrates a helpful paradigm for the control over acid-base bifunctional catalysts derived from LDH precursors, which will be of use in heterogeneous acid-base catalytic reactions.

This work was supported by the National Key R&D Program (2017YFA0206804), the National Natural Science Foundation of China (21871021, 21521005), and the Fundamental Research Funds for the Central Universities (buctylkxj01, XK1802-6).