Catalytic Conversion of Ethyl Levulinate to γ-Valerolactone Under Mild Conditions over Zr-Beta Acidic Zeolite Prepared by Hydrothermal Method

Abstract

As an important biomass resource, γ-valerolactone (GVL) shows considerable potential for applications in biofuel production, organic synthesis, polymer, and food industries. Herein, an effective method was presented for synthesizing GVL through the catalytic transfer hydrogenation (CTH) of ethyl levulinate (EL) under mild conditions. Using isopropanol as a hydrogen donor, a 100% conversion of ethyl levulinate and an 88.7% yield of GVL were achieved over 2%Zr-Beta-7d catalyst at 110 °C for 8 h. The acidic sites of synthesized Zr-Beta via hydrothermal methods easily adjusted and offered high catalytic activity and selectivity. The Lewis (L) acid sites on the zeolite serve as the active centers for the conversion of EL. Characterization results indicate that the amounts of L acid sites on Zr-Beta increased with the Zr content and crystallization time rose, thus enhancing the selectivity for GVL. Additionally, the influences of catalyst dosage, reaction temperature, and time on catalytic performance are studied, as well as calculations of kinetic parameters such as reaction rate constants and activation energies. The 2%Zr-Beta-7d catalyst retains its high performance after five cycles. The current research may present an efficient approach for the conversion of EL to GVL under mild conditions.

1. Introduction

During the rapid development of the global economy, a large amount of non-renewable fossil energy, such as coal and oil, has been rapidly consumed, which has caused increasingly prominent environmental problems and resource shortages. Therefore, it is urgent to find renewable resources [1,2]. Biomass is a kind of renewable organic carbon resource with carbon-neutral advantages that can achieve good ecological balance. Biomass conversion to target products usually requires biomass platform molecules as a bridge, which can be obtained through chemical catalysis such as alcohols, aldehydes, acids, ester, etc. [3,4,5,6,7,8]. Among these compounds, GVL is considered to be one of the most promising bio-based platform compounds, which can be used to produce important chemicals that partly replace fossil energy [9,10,11,12]. Therefore, GVL is of great significance for the development and utilization of bio-based resources.

GVL can be generated from a variety of biomass and its derivatives, such as cellulose [13,14,15,16], hemicellulose [17], fructose [18,19], xylose [20,21], and levulinate. Hydrochloric acid, solid acid [14], and precious metal catalysts [19,22] are used in the direct conversion of cellulose or hemicellulose to GVL. But these catalytic processes are usually complex and cannot achieve conversion in one step. Furthermore, the reaction selectivity is not high using cellulose or hemicellulose as substrate. The conversion of EL to GVL has attracted wide concern [23,24] because this reaction avoids the separation of intermediate products, reduces the production cost, and thus has a broad application foreground.

The preparation of GVL from EL can be divided into direct hydrogenation and catalytic transfer hydrogenation according to the hydrogen sources [25,26]. Chen et al. [27] reported a 100% conversion of EL and a 99% yield of GVL by 1%Pt/ZSM-35 catalyst at 6 MPa H₂ and 200 °C in ethanol solvent. When H₂ was used as a hydrogen source, the reaction conditions were harsh and the safety was low [28,29]. Based on the Meerwein–Ponndorf–Verley (MPV) reaction [30], GVL can be prepared by CTH of EL with alcohol as a hydrogen donor (Scheme 1). This process was under mild reaction conditions and exhibited high selectivity [31,32]. At present, precious metal catalysts (including Pd, Au, Ru, etc.) have been reported [33,34,35,36,37]. However, considering the economic feasibility, the development of non-precious metal catalysts has been widely considered [38,39]. Yang et al. [40] used Raney^®Ni catalyst and 2-propanol as a solvent, and the yield of GVL exceeded 90% at room temperature. However, the preparation process of the catalyst was complicated to preserve, and the performance was unstable. Liu et al. [41] reported that the conversion of EL was 91.6% and the yield of GVL was 81.8% for the NiCu-0.67 bimetallic catalyst in 2-propanol solvent at 220 °C for 6 h. In addition, Zr-based catalysts also have excellent catalytic performance in MPV reactions. Zhan et al. [42] synthesized the ZrO₂ catalyst with an abundance L acid sites, and the highest yield of GVL was 88.1% at 170 °C for 6 h. Yun et al. [43] designed a Zr metal–organic framework (Zr-MOF) with a yield of 98% GVL at 200 °C in 2-propanol solvent. The above-mentioned catalytic process has achieved a high yield of GVL. However, high reaction temperatures are required for these systems [42,44,45]. It is of great significance to develop a low-temperature and high-efficiency catalytic system suitable for this process.

Beta zeolite has the advantages of uniform pore structure, unique crystal structure, large specific surface area, and adjustable acidity, etc. [46]. As a typical L acid catalyst, Zr-Beta is generally used in MPV reduction reaction. To achieve the conversion of EL to GVL under mild conditions, the advantages of Zr species and Beta zeolite should be fully combined. Herein, the Zr-Beta zeolite synthesized by hydrothermal method was reported to efficiently catalyze the conversion of EL to GVL under mild conditions. A 100% conversion of ethyl levulinate and an 88.7% yield of GVL were achieved using isopropanol as a hydrogen donor. The Zr species were successfully introduced into the skeleton, and the L acid sites on the Zr-Beta zeolite were proved to serve as the active centers. The catalyst exhibited good recyclability. The kinetic experiments displayed that the CTH reaction was the first-order kinetic model.

2. Results and Discussion

2.1. Preparation and Characteristic

Zr-Beta zeolites with different Zr content and crystallization time were prepared by hydrothermal method. To reveal the effects of Zr content and crystallization time on the structure of xZr-Beta-yd catalyst, a variety of different characteristics were investigated. Figure 1a showed the N₂ adsorption and desorption isotherms of Zr-Beta zeolite with different Zr content. All the samples except for 3%Zr-Beta-7d presented type I adsorption and desorption isotherms, indicative of a typical micropore structure. The 3%Zr-Beta-7d samples showed mixed type I and type IV adsorption and desorption isotherms with an H4 hysteresis ring, indicating the part presence of a mesoporous structure [47]. It should be noted that H-Beta zeolite showed similar adsorption and desorption isotherms with the specific surface area of 416 m²/g.

Table 1. Physicochemical properties of samples with different zirconium content.

Sample *	SBET (m2/g)	Vt (cm3/g)	Vmicro (cm3/g)	Vmeso (cm3/g)
H-Beta-7d	416	0.434	0.357	0.077
0.5%Zr-Beta-7d	465	0.254	0.205	0.049
1%Zr-Beta-7d	463	0.260	0.207	0.053
2%Zr-Beta-7d	462	0.267	0.214	0.053
3%Zr-Beta-7d	451	0.330	0.190	0.140

* S_BET = the total surface area, V_t = the total pore volume, V_micro = the micropore volume, V_meso = the mesopore volume.

listed the physicochemical properties of Zr-Beta-7d samples with different Zr content. With the increase in Zr content, the specific surface area of the sample slightly decreased, the total pore volume gradually increased, and the mesoporous pore volume increased from 0.049 to 0.140 m²/g. The results showed that the increase in Zr content had an important effect on the pore structure of the zeolite, resulting in a partly mesoporous structure.

Furthermore, N₂ adsorption and desorption isotherms of 2%Zr-Beta samples with different crystallization times were shown in Figure 1b. All samples had similar type I adsorption and desorption isotherms, which were typical of microporous materials. With the extension of crystallization time, the specific surface area and micropore volume of 2%Zr-Beta samples gradually increased, while the mesoporous pore volume gradually decreased (Table S1), indicating that the micropore characteristics were enhanced with the extension of crystallization time, and the crystal structure gradually became complete.

The effect of Zr content and preparation time on crystallization was investigated using X-ray powder diffraction (XRD) characteristics. All the Zr-Beta-7d samples exhibited typical BEA topological structure (Figure 2a) [48]. When Zr content increased from 0.5% to 3%, the characteristic peak intensities (7.9° and 22.6°) were weakened, indicating that the crystallization rate of Zr-Beta-7d decreased gradually with the increase in Zr content. The characteristic peaks of 7.9° and 22.6° were found for both the H-Beta-7d and Zr-Beta-7d samples. These above results indicated that the zeolite skeleton basically did not change after the incorporation of Zr species into Beta zeolite. In addition, crystallization time was an important factor affecting the crystal structure and atom state of the zeolite. With the extension of crystallization time (Figure 2b), the characteristic peak intensity of the sample gradually enhanced, indicating that the crystallinity of the sample gradually increased. When the crystallization time exceeded 20 days, the characteristic peak intensity of the sample almost no longer changed, which was possible because the silicate in the gel would be consumed with the extension of the crystallization time. On the other hand, the zeolite generated under hydrothermal conditions would gradually dissolve with the extension of crystallization time and continued the crystallization process as a new nucleus [49].

The morphology structure of 2%Zr-Beta-7d was characterized by scanning electron microscopy (SEM) and transmission electron microscopy (TEM), as shown in Figure 3. The morphology of 2%Zr-Beta-7d was the capped square bipyramid (Figure 3a,b), which was in accordance with the characteristic morphology of Beta zeolite [31,32]. In addition, no obvious ZrO₂ particles were observed, indicative of a highly dispersed form of Zr element. Furtherly, EDS results confirmed that O, Si, and Zr elements were uniformly dispersed in Zr-Beta zeolite (Figure 3c).

Then, the chemical states of Zr, Si, and O elements in 2%Zr-Beta-7d were analyzed by X-ray photoelectron spectroscopy (XPS). It could be seen from the full spectrum of 2%Zr-Beta-7d that the sample contained Zr, O, and Si elements (Figure 4a). On the other hand, it also showed that Zr had been successfully introduced into zeolite. The binding energy of Si 2p in the fine spectrum of Si 2p was 104.1 eV (Figure 4b), corresponding to the Si⁴⁺ of the zeolite skeleton. In the fine spectrum of O 1s (Figure 4c), the peak shape of O 1s was symmetrical and the binding energy was 533.6 eV, corresponding to O²⁻, in the Zr 3d fine spectrum (Figure 4d), the binding energies of Zr 3d were 183.5 eV and 185.8 eV, and the corresponding double peaks were caused by the 3d_3/2 and 3d_5/2 spin orbits of Zr⁴⁺ [50].

The crystallization process of Zr-Beta had a significant effect on the L acid site in the catalyst skeleton. The types of acid sites and acid amount of Zr-Beta samples were analyzed by pyridine Fourier transform infrared spectrometer (Py-FTIR). All samples had obvious absorption peaks at 1446, 1490, and 1608 cm⁻¹, and the small absorption peaks at 1638 cm⁻¹ and 1545 cm⁻¹ could be assigned to the pyridine-adsorbed on Brønsted (B) acid sites (Figure 5). These results mean very small amount of B acid sites existed for the Zr modified Beta zeolite, which was consistent with reported work [42,51,52,53]. In addition, the absorption peak at 1490 cm⁻¹ was mainly attributed to the characteristic absorption peak of the interaction between pyridine and Zr-Beta at the L or B acid sites [54], indicating that the above samples primarily contained L acid sites. The acid amount was further calculated according to the peak area (Table S2). The L acid sites of Zr-Beta-7d significantly increased from 31.56 to 93.63 µmol/g with the Zr content increasing from 0.5% to 3%, indicating that the introduction of Zr species facilitated to form the L acid sites (Figure 5a). On the other hand, the L acid sites of 2%Zr-Beta increased from 64.81 to 93.19 µmol/g when the crystallization time was gradually expanded from 3 to 25 days (Figure 5b), indicating that the longer crystallization time was beneficial to generate the L acid sites of 2%Zr-Beta samples.

2.2. Effect of Zr Content and Crystallization Time on Catalytic Performance

For the catalytic conversion of EL to GVL (Figure 6a), H-Beta exhibited a poor activity with a low conversion of 35.2% under mild conditions. As for Zr-Beta catalysts, it showed the significant improvement of EL conversion and GVL yield. When Zr content was 0.5%, the EL substrate cannot be completely transformed, which was due to the low Zr content related to fewer L acid sites for 0.5%Zr-Beta-7d catalyst. With the gradual increase in Zr content, the EL was completely transformed, and the yield of GVL tended to be stable after the obvious increase. When the Zr content was 2%, the highest yield of GVL reached up to 88.7% (Table S3). Thus, the L acid sites with different Zr content had a great influence on the transformation of EL to GVL.

Furtherly, for 2%Zr content, the EL kept complete transformation, and the yield of GVL was also gradually increased from 84.8% to 88.7% as the crystallization time gradually increased from 3 to 7 days (Table S4). When the crystallization time exceeded 7 days, the yield of GVL hardly increased (Figure 6b). The above results indicated that Zr-Beta with a certain crystallization time had the appropriate amount of L acid sites, thus showing good catalytic performance in the conversion of EL to GVL.

2.3. Effect of Reaction Conditions on Conversion of EL to GVL

Solvent as a hydrogen donor played a key role in this reaction. As shown in Figure 7a, the influence of different alcohol solvents on the conversion of EL was explored. When methanol and ethanol were selected as the reaction solvents, the conversion of EL was relatively low, which was due to the high reduction potential and the poor performance as hydrogen donors for these two solvents. Generally, the reduction potential of secondary alcohols was lower than that of primary alcohols, so when isopropanol was used as a hydrogen donor, the conversion of EL and the yield of GVL were higher than that of n-propanol. Moreover, the reaction effect was further reduced when the butanol was used as a hydrogen donor, maybe due to the long carbon chain and large spatial structure. Therefore, EL was completely converted with 2-propanol as a hydrogen donor, and the yield of GVL was 88.7%, which was the most efficient hydrogen donor in the scope of the current investigation (Table S5).

As shown in Figure 7b, the yield of GVL increased from 83.7% to 88.7% when the catalyst dosage gradually increased from 0.1 to 0.2 g (Table S6). However, the yield of GVL hardly increased as the mass of catalyst continued to increase. Figure 7c showed the effect of reaction temperature on the conversion of EL to GVL. When the reaction was carried out at 110–130 °C, the yield of GVL increased from 86.0% to 88.7% (Table S7). However, the yield of GVL decreased slightly once the reaction temperature continued to increase. The effect of reaction time was shown in Figure 7d. The conversion and the yield, respectively, increased from 48.8% to 100% and 20.8% to 88.7% with the extension of reaction time from 0.5 to 8 h (Table S8). The yield of GVL remained constant, indicative of the equilibrium state for the reaction at longer time (over 8 h).

For heterogeneous catalysis, catalyst stability is an important reference. To evaluate the stability, the 2%Zr-Beta-7d after the reaction was separated and recycled (Figure 8a). The reaction substrate was added in equal proportion to the mass of the actual catalyst each time. The yield of GVL did not decrease significantly after 5 cycles of catalytic reactions (Table S9). Meanwhile, ICP-OES was used to analyze the content of Zr in the reaction solution after the first and fifth cycles. The content of Zr in the above two reaction solutions was lower than 1.0 ppm, indicating no loss of active species for the catalyst during cycle experiments. In addition, the XRD pattern of 2%Zr-Beta-7d catalyst was provided before and after the cycle (Figure 8b). The characteristic peak intensity of the catalyst was basically unchanged after five cycles, indicating the remained crystal structure of 2%Zr-Beta-7d, and the topological structure of Beta zeolite was still retained. The N₂ adsorption and desorption isotherms of 2%Zr-Beta-7d before and after cycling were shown in Figure 8c, and the catalysts both showed type I adsorption and desorption isotherms, indicating that the zeolite structure of 2%Zr-Beta-7d does not change significantly after recycling. On the other hand, the FT-IR spectrum was also studied for the reused catalyst (Figure 8d). The signal of major functional groups does not change, but a relatively obvious peak appeared at 2931 cm⁻¹, which belonged to the stretching vibration peak of C-H, indicating that some organic carbon might be adsorbed on the surface of the catalyst.

SEM and TEM images of the reused 2%Zr-Beta-7d catalyst were shown in Figure S1. After cycling five times, the catalyst still retained the original relatively regular geometry, and its fine structure did not change significantly (Figure S1a). In addition, the reused catalyst after cycle still retained the capped square bipyramid morphology, and there was no significant change in its morphology compared with that of fresh 2%Zr-Beta-7d (Figure S1b). Meanwhile, EDS diagrams showed that the Zr element was still evenly distributed (Figure S1c). In conclusion, the above experimental and characterization results indicated that 2%Zr-Beta-7d had high stability under the current reaction.

2.4. Kinetic Experiments

The kinetics experiments of EL conversion to GVL by 2%Zr-Beta-7d catalyst were investigated. To avoid the interference of mass transfer resistance, the reaction was carried out at a low conversion of EL (less than 30%). The relationship of concentration versus time confirmed the reaction was the first-order kinetic model for the 2%Zr-Beta-7d catalyst at 110–130 °C (Figure 9). In addition, the activation energy (E_a) was 77.33 kJ/mol for the conversion of EL to GVL catalyzed by 2%Zr-Beta-7d catalyst.

3. Materials and Methods

3.1. Chemicals

EL (99%), 2-propanol (99.9%), GVL (98%), pyridine (≥99%), tetraethyl orthosilicate (TEOS, AR grade), zirconium oxychloride octahydrate (ZrOCl₂·8H₂O 99%), naphthalene (AR grade), and tetraethylammonium hydroxide (TEAOH, AR grade) were from Aladdin Chemical Reagents (Shanghai, China). Si-Beta (SiO₂/Al₂O₃ = 12.5) was purchased from Nankai University Catalyst Co., LTD (Tianjin, China). Hydrofluoric acid (40 wt%) and nitric acid (65 wt%) were purchased from Zhengzhou Xihua Co., LTD (Zhengzhou, China). All commercial reagents, unless otherwise noted, were used without further purification treatment.

3.2. Zr-Beta Zeolite Prepared by Hydrothermal Methods

Zr-Beta zeolite with different Zr content and crystallization time were synthesized by hydrothermal method in a fluorine-containing system using Si-Beta as crystal seed. TEOS (34.95 g) was slowly dropwise added to the stirred TEAOH (52.9 g, 25 wt% aqueous solution). After stirring for 1.5 h, a certain amount of SnCl₄·5H₂O aqueous solution was added, and the TEOS was completely hydrolyzed by stirring for 12 h. Then, the mixture was heated in a water bath at 80 °C until the ethanol produced by hydrolysis was removed from the system. HF (0.54 g) was added to the solution, stirred until the gel was transferred to the stainless steel reactor, and then the mixed solution of Zr-Beta seed (0.4 g) and ultrasonic deionized water (3.0 g) was dropwise added to the gel. Finally, the gel is transferred to the stainless steel crystallization reactor and placed in the oven at 140 °C. After a certain time, it is removed and cooled rapidly to stop crystallization. The solid was collected, washed, and dried overnight, and calcined at 550 °C in a Muffle furnace for 8 h to remove the templating agent. The final zeolite was denoted as xZr-Beta-yd (x represents Zr content, and yd represents different crystallization days).

3.3. Characteristics of Catalysts

To test the crystal structure, XRD was carried out by Bruker D8 Advance (Bruker Technology LTD, Wolzbach, Germany) solid powder diffractometer. Cu-Kα (λ = 0.1548 nm) was used as the ray source. The tube voltage and tube current were 45 kV and 40 mA, respectively. The 2θ scanning range is 5° to 80°. N₂ physical absorption and desorption were collected using the American Micrometrics ASAP 2460 physical adsorption instrument (McMuratic Instruments Co., LTD, Norcross, Georgia, USA). Before the test, the sample was degassed at 150 °C for 12 h under vacuum conditions, then adsorbed high-purity N₂ and performed adsorption–desorption test under 77 K liquid nitrogen conditions. To analyze the coordination state of Zr in Zr-Beta zeolites, the Cary series UV-vis-NIR spectrophotometer (Agilent Technology Co., LTD, Palo Alto, California, USA) was used to record data in the range of 200 nm to 600 nm. XPS was used to test the sample on the Thermo ESCALAB 250Xi photoelectron spectrometer (Thermo Fisher Technologies LTD, Waltham, Massachusetts, USA). The test parameters were as follows: Al-Kα X-ray was used as the excitation source, the voltage was 15.0 kV, and the indoor pressure was less than 1 × 10⁻⁹ mBar; the binding energy of the sample elements was corrected based on C1s = 284.80 eV. The sample should be vacuum-dried at 80 °C for 12 h before testing. FT-IR was used to test the sample using a German Bruker Tensor II Fourier infrared spectrometer (Bruker Technology LTD, Wolzbach, Germany) with a scanning wavelength range of 4000 to 400 cm⁻¹. The FTIR spectroscopy of pyridine adsorption was performed on the Bruker Tensor II spectrometer. The sample (20 mg) was first pretreated under vacuum at 450 °C for 3 h, and then a background reference (FTIR spectrum of the sample under vacuum) was collected at room temperature. Pyridine was then adsorbed until saturated, followed by being heated to different temperatures under vacuum and kept for 0.5 h for desorption. After cooling to room temperature, the spectrum at different desorption temperatures was collected and scanned with a resolution of 4 cm⁻¹. The acid content is calculated according to the following formula [55]:

$$\mathrm{Acid}\mathrm{site}\mathrm{density}(\mathsf{\mu }\mathrm{mol}/\mathrm{g})={\displaystyle \frac{\mathrm{Integrated}\mathrm{peak}\mathrm{area}({\mathrm{c}\mathrm{m}}^{-1})}{\mathrm{E}(\mathrm{cm}/\mathsf{\mu }\mathrm{mol})}}\times {\displaystyle \frac{{\mathrm{A}}_{\mathrm{c}\mathrm{s}}({\mathrm{c}\mathrm{m}}^2)}{\mathrm{m}\left(\mathrm{g}\right)}}$$

where acid site density represented the acid density of L, integrated peak area represented the area of the absorption peak of the interaction between the acid site of L and the pyridine molecule, E represented the extinction molar coefficient, A_cs represented the cross-sectional area of this circular sheet, and m represented the mass of this circular sheet.

3.4. Catalytic Reaction and Analysis

A Teflon-lined stainless steel autoclave (15 mL) was used as the reactor. First, a certain amount of EL was dissolved in 5 mL solvent and then transferred to the above reactor. A certain amount of catalyst was added, and then the reactor was sealed, stirred, and heated to a set temperature with a specified program. After the reaction, the reactor was quickly cooled in ice water, and cooled to room temperature, with naphthalene as the internal standard added. The mixture was analyzed qualitatively and quantitatively by Agilent 8860/5977B gas chromatography-mass spectrometry (GC-MS) (Agilent Technology Co., LTD, Palo Alto, California, USA). The gas chromatographic column was OV-1701 capillary column (30 m × 320 µm × 0.25 µm), the inlet temperature was set at 240 °C, and the initial column temperature was 80 °C for 5 min. The temperature was raised to 180 °C at 15 °C/min and kept for 15 min. The conversion of EL and the yield of GVL are calculated according to the following formula:

$$\mathrm{Conversion}\mathrm{of}\mathrm{EL}\left(\%\right):\mathrm{C}={\displaystyle \frac{\mathrm{mole}\mathrm{of}\mathrm{the}\mathrm{consumed}\mathrm{EL}}{\mathrm{mole}\mathrm{of}\mathrm{the}\mathrm{initial}\mathrm{EL}}}\times 100\%$$

$$\mathrm{Yield}\mathrm{of}\mathrm{GVL}\left(\%\right):\mathrm{Y}={\displaystyle \frac{\mathrm{mole}\mathrm{of}\mathrm{the}\mathrm{generated}\mathrm{GVL}}{\mathrm{mole}\mathrm{of}\mathrm{the}\mathrm{initial}\mathrm{EL}}}\times 100\%$$

$$\mathrm{Selectivity}\mathrm{of}\mathrm{GVL}\left(\%\right):\mathrm{S}={\displaystyle \frac{\mathrm{Y}}{\mathrm{C}}}\times 100\%$$

4. Conclusions

In summary, Zr-Beta zeolites with different Zr content and crystallization time were synthesized by hydrothermal method, and catalytic transfer hydrogenation of EL to GVL was achieved by the Zr-Beta zeolite under mild conditions. The L acid sites on the catalyst acted as the active center for the catalytic conversion, and it evidently increased with the increase in Zr content and crystallization time and showed good catalytic performance when increasing to a certain amount. The hydrogen donor also had a great influence on the catalytic performance. Compared with other alcohol solvents, 2-propanol had the best effect when it was used as a hydrogen donor. EL was converted completely and the yield of GVL was 88.7% over 2%Zr-Beta-7d using 2-propanol solvent at 110 °C for 8 h. The catalyst had excellent stability and maintained high catalytic activity after five cycles. In addition, the CTH reaction was proved to be the first-order kinetic model and the activation energy (E_a) was 77.33 kJ/mol.