F⁻@d4r, a New Type of Acidic Catalytic Site in Zeolite

Abstract

As a solid acid, zeolite has been widely used in many fields such as tail gas treatment, petrochemical engineering, and the fine chemical industry. F⁻ has been widely used in the synthesis of pure silica or high silica zeolites. To balance the charge of organic structure directing agents (OSDA), F⁻ is often found located at the center of the double-4-rings (d4r) of the as-made zeolites. During calcination, fluorine ion is removed with the OSDA. We screened a series of composition building units and found that d4r is capable to retain F⁻ in zeolite structure. We introduce the F⁻ back after the calcination to create an unprecedented type of acid site, i.e., F⁻@d4r in pure silica zeolites ITQ-12. The F⁻@d4r species is thermal stable up to 300 °C. ITQ-12 with F⁻@d4r shows substantial catalytic activity in Biginelli reaction. Furthermore, the catalytic performance is proved to be positive correlated with the presence of F⁻@d4r, indicating the mild acid catalytic property of F⁻@d4r.

1. Introduction

Zeolites, a group of inorganic crystalline materials with micropores, were primarily synthesized in alkaline conditions. By 1978, Flanigen and Patton [1] introduced F⁻ anion as the mineralizing agent into the synthesis of silica-rich zeolites, and this method was known as the “fluoride route”. High pH value can be avoided with the presence of an F⁻ anion in the synthesis gel, enabling the incorporation of transition metal into the zeolite framework. The resulting products exhibit fewer framework defects [2,3], and well-defined shapes [4]. The reason is that the positive charge of the organic structure-directing agent (OSDA) is balanced by F⁻ anion instead of the defects in the framework [5]. During the last three decades, the fluorine route has been developed extensively, and has led to the synthesis of silica-rich zeolites [6], as well as a number of zeolites with novel structures [7,8,9,10].

It is widely noticed that zeolites could be synthesized in fluoride media. F⁻ ion is not only acting as a mineralizer but also participating in product formation. The F⁻ ion located in small cages such as double four-ring units (d4r), bea [3], rth [11], non [12], [4¹5²6²], etc., can be investigated by ¹⁹F NMR spectroscopy or X-ray diffraction [10,13,14]. Furthermore, as-synthesized zeolites are calcined to remove the organic component, F⁻ ion is removed as well [15], leaving zeolites with open cages and pores. Liu et al. [16] reported that F⁻ ion could ion-exchange. Due to the size effect, F⁻ ions in non-d4r cages are not stable enough that they could be ion-exchanged with ammonia solution. Whereas the F⁻@d4r (F⁻ inside of d4r) is stable with the ion exchange. However, they perform the ion-exchange with the OSDA inside the pores. Therefore, the resulting product is not porous material. Generally speaking, OSDA decomposed over 400 °C and left the empty channels or cages, the F⁻ ion leave below the temperature. Thus, the F⁻@d4r composite materials are not yet investigated along with the zeolite as a porous material.

In pure-silica zeolites with d4r, the bond angle of Si-O-Si is about 148°. The “bond angle” of Si-Si-Si in the corner of the cube is 90°. The mismatch of the bond angle leads a tensile force. Thus, synthesis of pure-silica zeolites with d4r is difficult at high pH value [17]. In the fluorine route, F⁻ ions contribute to formation and stabilization of d4r units. The d4r framework remains stable after the F⁻ ions removed, even if the tensile forces still exist. Analysis with DFT and Kohn-Sham theory indicated that F⁻ ions in the center of silica d4r cage contribute to stability of d4r, which cannot be achieved by other halogen ions [18]. According to the calculation, F⁻ ions form hypervalent bonds with the surrounding eight silicon atoms. The coordination number is 8, the bond order is 4, while the bond is highly delocalized.

ITQ-12 is a pure silica zeolite synthesized with trimethylimidazolium salt as OSDA reported by A. Corma et al., in 2003 [17]. IZA (Inernational Zeolite Association) three letter code is ITW [19]. This framework has two-dimensional eight-ring channels, and the size of eight-ring channel in the (100) direction is ellipse shape (2.4 × 5.3 Å) while (001) direction is round (3.8 × 4.1 Å). The result of structure elucidation indicates that the OSDA is in [4⁴5⁴6⁴8⁴] cages, and the cages are connected by double-four rings. Therefore, there are a large number of double-four rings in the framework.

In this manuscript, pure-silica double four-ring units (d4r) are proved to be able to stabilize F⁻ ion. The properties of a F⁻@d4r zeolites, ITQ-12-F are studied in detail via a series of characterization methods. The catalytic performance of this ITQ-12-F is also evaluated in comparison to commercial HY zeolite in Biginelli reaction [20,21,22].

2. Results

2.1. Characterization of Catalysts

In this work, several F⁻ species, i.e., F⁻ in zeolite channel, F⁻ trapped in the defects, F⁻ in the small cages, have been recreated by the F⁻ treatment in the “empty zeolites” and evaluated by the thermal analysis. However, most F⁻ species are not thermally stable enough to be isolated from others but F⁻@d4r (d4r with F⁻ ion inside). The F⁻ ions have been reintroduced into d4r of calcined zeolites ITQ-12, and it could be stable up to 300 °C. The preparation is shown in Scheme 1. To understand the nature of the F⁻@d4r, we intend to choose the pure silica zeolite to avoid the effect of heteroatoms. If the F⁻ reintroduction brings the acid site, we can safely conclude that the acidic property originates from F⁻@d4r.

Firstly, we are trying to determine how much F⁻ species can retain by the pure silica zeolite framework at different temperatures. To gain knowledge about the affinity of F⁻ ion by the channels, defects, small cages, i.e., bea, [4¹5²6²], d4r. Silicalite-1, SSZ-74 [23], Beta [6], ITQ-12 [17] are prepared (denoted as Silicalite-1-as, SSZ-74-as, Beta-as, ITQ-12-as) and calcined. Then the samples are treated by NH₄F (marked as Silicalite-1-F, SSZ-74-F, Beta-F, ITQ-12-F). The subsequent thermal treatment leads to a series of samples named Silicalite-1-F-X, SSZ-74-F-X, Beta-F-X, ITQ-12-F-X, where the X denotes thermal treatment temperature. Figure S1 shows the comparison of XRD spectra of four kinds of molecular sieves before and after fluorine treatment in NH₄F-methanol solution. The peak differences probably raised from the introduction of F⁻ ion into the structure. Furthermore, the average intensity of the XRD shows no obvious decrease indicating the crystallinity of the molecular sieves retains after the treatment. ITQ-12-F The corrosion degree of zeolite can be inferred by SEM images of the samples (Figure S2). The crystal surface of ITQ-12 remains intact, indicating the subtle corrosion degree of the F⁻ treatment. While Beta and Silicate-1 demonstrate substantial surface corrosion.

F⁻ post-treatment is a common method to create mesoporous in the zeolite, especially HF, HF₂^- are major species to corrode SiO₂ [24]. Hence the proper concentration and ion-exchange temperature are necessary to avoid the creation of mesopore. The samples are firstly treated by NH₄F solution of methanol and then are heated at different temperatures. The F⁻ contents of the resulting samples are evaluated by the F⁻ ion-selective electrode, as seen in Figure 1. (a) Fluoride contents of Beta, Silicate-1, SSZ-74, ITQ-12 after heated at different temperature; (b) Fluoride contents of ITQ-12 F⁻ loaded in different concentrations of NH₄F-methanol solution. The Silicalite-1 and SSZ-74 quickly lose their all F⁻ contents below 200 °C. At the same time, Beta and ITQ-12 retain more than 1.5%wt up to 400 °C. To obtain ITQ-12-F with different fluoride contents, ITQ-12 is loaded F⁻ ion in various concentrations of NH₄F-methanol solution. The result indicated that the F⁻ content shows positive correlation with the NH₄F concentration and reaches the maximum value when the concentration of NH₄F is equal or greater than 0.1mol/L (Figure 1. (a) Fluoride contents of Beta, Silicate-1, SSZ-74, ITQ-12 after heated at different temperature; (b) Fluoride contents of ITQ-12 F⁻ loaded in different concentrations of NH₄F-methanol solution). ¹⁹F MAS NMR is performed to understand the F⁻ species retained by Beta and ITQ-12, as seen in Figure 2. The ¹⁹F-MAS-NMR of (a) Beta, Beta-F and Beta-F-300; (b) ITQ-12, ITQ-12-F-calcined, ITQ-12-F-300 and ITQ-12-F-350. a. As-synthesized zeolite beta, the resonances at −58 and −70 ppm are assigned to the bea cage. The small resonances at −38 ppm indicated that a small amount of d4r units derived from polymorph C (BEC) naturally occurring as intergrowth. After calcination and NH₄F treatment, the F⁻ ion in the small cages left as the disappearance of −38, −70 ppm. The signal at about −120 ppm attributed to the dissociated F⁻ ions in zeolite channels [24]. The slight resonance at −58 ppm indicates the F⁻ in bea restored. Then the thermal treatment at 300 °C eliminated all the weak “trapped” F⁻. The dissociated F⁻ ion are partially removed while the signal of d4r trapped F⁻ (denoted as F⁻@d4r) in BEC intergrowth is drastically enhanced. Implying the d4r bearing zeolite is a better example of current research. Therefore, we turn our focus on the d4r zeolite ITQ-12. Similar treatments are performed for the ITQ-12 (Figure 2. The ¹⁹F-MAS-NMR of (a) Beta, Beta-F and Beta-F-300; (b) ITQ-12, ITQ-12-F-calcined, ITQ-12-F-300 and ITQ-12-F-350. After calcination and NH₄F treatment, F⁻@d4r (−38 ppm) decomposes completely. Reintroduced flourine principally bond with Si and partially dissociate in zeolite channels as the form of F⁻ ions (−122 ppm), instead of regenerating F⁻@d4r. The signal at −150 ppm is attributed to SiO_3/2F species, while −106 and −194 ppm are attributed to other F⁻ species bonding with Si [24]. The thermal treatment at appropriate temperature (300 °C) causes the most F⁻ trapped by d4r, leaving only a tiny amount of SiF₅^- species (−129 ppm). The higher temperature of 350 °C destroys most of F⁻@d4r and lead fluorine to dissociate in channels or bond with Si. Hence, we know the F⁻@d4r is stable up to 300 °C, enough for the most liquid phase catalytic reactions. ²⁹Si MAS NMR is performed to investigate the Si species on zeolites, as seen in Figure S5. Si species has little changes after the process of calcination and NH₄F treatment.

If the F⁻ anion is encapsulated within the charge-neutral pure silica zeolites, a cation must present to compensate the charge. The NH₄⁺ is introduced at the same time when F⁻ is introduced. Similar to the usual step to expose the acidic site of zeolite after NH₄⁺ ion exchange, the NH₃ escapes during the 300 °C thermal treatment leaving the proton free or attached to the framework oxygen. Therefore, the Bronsted acidic site is suggested for an F⁻@d4r. NH₃-TPD curve shows a peak centered at 100 °C corresponding to the NH₄- F⁻@d4r. The monotonic increasing of signal suggests the losing F⁻ and the band at 550 °C indicated the considerable losing of F⁻ (Figure S3). The texture property experiments of ITQ-12 and ITQ-12-F-300 show a typical I type isotherm. ITQ-12-F-300 exhibits a slight decrease compared to the ITQ-12 (550 °C calcined) samples (BET surface area, 528 m²/g vs. 547 m²/g, isotherm as seen in Figure S4). It indicates the open pore structure. Table S1 illustrates a measurable increase of zeta potential (absolute value) and a slight decrease of surface energy after reintroducing F⁻ ions. This can be owed to the appearance of surface charge, which balances the acid sites on surface of zeolites.

Furthermore, due to the strictly 1 to 1 ratio of F⁻ and d4r, the maximum species of F⁻@d4r is easy to calculate. First, we can calculate the ratio of Si atom and d4r in a unit cell, which is 12 to 1. Next, Si to F ratio is 12 to 1. So, the maximum F⁻ species contents of F⁻@d4r is 2.57% which is in line with the experimental data, i.e., 2.34%. The formula as following, where $M{r}_{\mathrm{SiO}2}$ and $M{r}_{\mathrm{F} \mathrm{ion}}$ are molecular weight of the SiO₂ and F⁻ ion, respectively.

$$\mathit{Theoretical} {\mathrm{F}}^{-} \mathit{content}=\frac{M{r}_{\mathrm{F} \mathrm{ion}}}{M{r}_{\mathrm{SiO}2}\times 12+M{r}_{\mathrm{F} \mathrm{ion}} }$$

Generally, the acidic properties of zeolites could be directly characterized by solid-state ¹H NMR. However, the OSDA in the channel was difficult to be completely removed, and the residual organic species would interfere with the NMR spectrum. Thus, we identified the acidities of F⁻@d4r zeolites by means of Biginelli reaction, a classic acid-catalyzed reaction.

2.2. Catalytic Performance

The Biginelli reaction can be catalyzed by Brønsted acids and/or Lewis acids. Scheme 2 illustrates the reaction formula. We try to use ITQ-12 and commercial HY zeolites (Si/Al ratio = 5.0) as catalysts for Biginelli reaction, urea, benzaldehyde, and ethyl acetoacetate as raw materials under the same conditions. The purified product was quantified by ¹H NMR (Figure S6). Comparing the yield of Biginelli reaction with different catalysts (Figure 3. Biginelli reaction yield of (a) various zeoilte catalysts and blank; (b) ITQ-12-F-300 with various F⁻ contents; (c) ITQ-12 treated in various temperature; (d) reusing ITQ-12-F (The reaction conditions are 80 °C for 10 h). (a) indicates that bare ITQ-12 shows an identical yield curve with blank reaction. That suggests the plain ITQ-12 sample is inactive in the reaction. After introducing F⁻ ion in the d4r in ITQ-12, the productivity of Biginelli reaction showed a significant increase to 47.88%. Though its catalytic performance did not bear comparison with commercial HY zeolite (Si/Al ratio = 5.0), the gap of their catalytic performance has narrowed considerably after F⁻ ion loaded. Thus, it can be inferred from the catalytic activity that F⁻@d4r zeolites possessed acidity, which might result from H⁺ ion, the charge-balancing cation of F⁻@d4r species.

Furthermore, we investigated the catalytic performance of F⁻@d4r zeolites with different F⁻ contents. The result indicated a positive correlation between F⁻ contents, and yield of product (Figure 3. Biginelli reaction yield of (a) various zeoilte catalysts and blank; (b) ITQ-12-F-300 with various F⁻ contents; (c) ITQ-12 treated in various temperature; (d) reusing ITQ-12-F (The reaction conditions are 80 °C for 10 h). b). Figure 3. Biginelli reaction yield of (a) various zeoilte catalysts and blank; (b) ITQ-12-F-300 with various F⁻ contents; (c) ITQ-12 treated in various temperature; (d) reusing ITQ-12-F (The reaction conditions are 80 °C for 10 h). (c) shows the “Λ shape” trend (first increasing and then decreasing) with the increasing of treatment temperature of catalysts. Firstly, the relation between F⁻ content and yield of the production still positive (increasing first and decreasing after 300 °C). Secondly, the treating temperature high than 300 °C leading the decomposition of F⁻@d4r. Consequently, the product yield decrease. The trend further confirm the conclusion, i.e., the F⁻@d4r is the real acidic catalytic center. That is in line with the previously stated facts with the experimental result of ¹⁹F MAS NMR (Figure S4b). Therefore, we can conclude that the catalytic activity originated from F⁻@d4r rather than other species of fluorine. The durability test shows the ITQ-12-F remain stable after at least 3 times of Biginelli reaction without the F content leaching (Figure 3. Biginelli reaction yield of (a) various zeoilte catalysts and blank; (b) ITQ-12-F-300 with various F⁻ contents; (c) ITQ-12 treated in various temperature; (d) reusing ITQ-12-F (The reaction conditions are 80 °C for 10 h).

3. Materials and Methods

3.1. Synthesis of Zeolites

ITQ-12 was synthesized following the recipe reported by Camblor et al. Fumed silica was used as the silica source, and homemade 1,2,3-trimethylimidazolium hydroxide (TMIOH) was used as the OSDA. Gels with the composition of SiO₂:0.5 TMIOH:0.52HF:7H₂O was placed in stainless-steel Teflon-lined autoclaves and heated at 175 °C for 7 days. Solids were recovered by filtration and washed. After dried overnight, zeolites were calcined in air at 600 °C for 6 h [1].

Pure SiO₂ zeolite-β was synthesized following the recipe reported by Camblor et al. [1]. Tetraethylorthosilicate (TEOS) was used as the silica source and tetraethylammonium hydroxide (TEAOH) was used as the OSDA [1]. Gels with the composition of 0.54NEt₄OH:0.54HF:SiO₂:7.25H₂O was placed in stainless-steel Teflon-lined autoclaves and heated at 160 °C for 39 h while being rotated at 60 rpm. After recovered by filtration, washed and dried overnight, solids were calcined in air at 550 °C for 6 h [1].

Pure silica SSZ-74 was synthesized following the recipe reported by Baerlocher et al. [23] Tetraethoxysilane was used as the silica source and 1,6-bis(N-methylpyrrolidinium)-hexane was used as the OSDA. Gels with the composition of 0.4ROH:0.26HF:0.8SiO₂:2.8H₂O was placed in stainless-steel Teflon-lined autoclaves and heated at 150 °C for 22 days. The solids were recovered by filtration, washed with demineralized water and ethanol, and finally dried overnight at 60 °C.

F⁻ ion was loaded into pure-silica zeolites by mixing calcined zeolites with 0.1 mol/L NH₄F in methanol [25]. After stirring at 65 °C for 12 h, the solid product was recovered by filtration and washed with water, then dried at 60 °C overnight. Finally, the solid was heated at 300 °C (for ITQ-12 and Beta) or 100 °C (for Silicate-1 and SSZ-74) for 6 h. To obtain F⁻@d4r with various F⁻ contents, concentrations of NH₄F-methanol solution were varied.

3.2. Catalytic Study

F⁻@d4r zeolites were activated in muffle furnace at 200 °C for 5 h. 0.2 g catalyst, ethyl acetoacetate (0.006 mol, 99%) purchased from Shanghai Aladdin Biochemical Technology Co., Ltd. (Shanghai, China), benzaldehyde (0.006 mol, 99%, Shanghai Aladdin Biochemical Technology Co., Ltd.) and urea (0.0072 mol, 99%) obtained from Sigma-Aldrich (Shanghai) Trading Co., Ltd. (Shanghai, China) in glacial acetic acid (30 mL, 99.8%, Shanghai Aladdin Biochemical Technology Co., Ltd.) were loaded into a three-necked flask fitted with a reflux condenser and a thermocouple. The mixture was treated by heating reflux at 80 °C for 10 h. After the reaction completed, the mixture was then cooled down to room temperature. The catalyst was removed by filtration and 60 mL icy water poured on the solution. The purified product was obtained by filtering the solid product and recrystallizing from ethanol.

3.3. Determination of F⁻ Contents

Fluoride ion content in ITQ-12 were determined by method of ion-selective electrode analysis. 20 mg ITQ-12 was placed in a nickel crucible with 0.2 g NaOH and heated at 550 °C for 20 min in the muffle furnace. Calcined zeolite was soaked in approximately 50 mL H₂O and transferred into 50 mL volumetric flask. 0.2 mL hydrochloric acid was then added to the solution and H₂O was added to dilute the solution to the tick mark. 10.0 mL solution with 1–2 drops of bromocresol purple indicator was transferred into 50 mL volumetric flask. The solution pH was adjusted with drop hydrochloric acid until the solution changed color from violet to yellow. 15 mL TISAB was added, and the solution was diluted with deionized water to 50 mL. The test solution was then transferred into polyethylene beakers and stirred. Fluoride ion-selective electrode and Ag-AgCl reference electrode was used to determine the electrode potential. The standard curve was established. Fluorine content was calculated according to the electrode potential and the standard curve.

3.4. Zeolites Characterization

The powder X-ray diffraction (PXRD) patterns were investigated by Bruker D2 PHASER with Cu-Kα radiation (λ = 1.54178 Å) at 30 kV, 10 mA (Figure S1). Scanning electron microscope (SEM) were measured on SU8010 microscopy (Figure S2). The measurement of ammonia temperature-programmed desorption (NH₃-TPD) was detected by a MFTP-3060 instrument (Figure S3). Before the measurement, 0.1 g of zeolite was placed in a slender quartz tube and then baked for 1 h at 300 °C under helium flow (40 mL/min). The zeolite was then cooled down to 30 °C. After adsorbing ammonia/helium (40 mL/min) at 100 °C for 0.5 h, helium (40 mL/min) was introduced for 5 min to purge the sample and remove ammonia physically adsorbed. NH₃-TPD measurement was then performed in 30–730 °C, heating at the rate of 10 °C/min with helium as carrier gas. The ammonia desorbed was recorded on a thermal conductivity detector. ¹⁹F MAS-NMR (Figure S4) and ²⁹Si MAS-NMR (Figure S5) were recorded on a Bruker AVANCE III 500 MHz Superconducting Fourier Transform Nuclear Magnetic Resonance Spectrometry. The ¹H NMR analysis of catalytic product was determined by Bruker advance III 400 MHz Nuclear Magnetic Resonance spectrometer (Figure S6). Zeta potentials of zeolites was measured by an EliteSizer Omni instrument (Brookhaven Instruments Corp.) at pH = 7 and surface energy was investigated using the KRÜSS DSA100 shape analyzer (Table S1).

4. Conclusions

In conclusion, we have successfully introduced F⁻ ion into the d4r in ITQ-12 by post F⁻ treatment while keeping the channel open. The F⁻@d4r species demonstrated a thermostability up to 300 °C, and it would generate Si-F bonds with the framework of zeolites at 350 °C. According to calculations, the maximum F⁻ species contents of F⁻@d4r is 2.57% which is in line with the experimental data, i.e., 2.34%. To investigate the acidity of F⁻@d4r zeolites, NH₃-TPD and ¹H NMR experiment were performed, but did not come to expected target. Thus, we identified the acidities of F⁻@d4r zeolites by means of Biginelli reaction, a classic acid-catalyzed reaction. The catalytic experiment reveals that F⁻@d4r allowed pure silica zeolites to gain acidity and acid-catalytic performance, and its acid-catalytic performance is positive correlated with the presence of F⁻@d4r. We do notobserve an obvious catalytic performance decay and F⁻ leaching in the reusability test.