Sustainable production of ethylene from bioethanol over hierarchical ZSM-5 nanosheets

Abstract

Hierarchical aluminosilicate nanosheets composed of an MFI structure with various Si/Al ratios have been successfully prepared via a one-pot hydrothermal process with the aid of tetrabutylphosphonium hydroxide (TBPOH) as the bifunctional structure-directing agent (SDA). The MFI zeolite nanosheets exhibit outstanding properties, such as an extremely high external surface area and appropriate acidic properties. To illustrate the beneficial effect of the hierarchical structure of the zeolite nanosheets towards green and sustainable catalytic conversion of renewable resources to high value-added chemicals, direct bioethanol dehydration to ethylene over Brønsted-acid MFI nanosheets has been studied from both experimental and theoretical points of view. Interestingly, the hierarchical structure strongly effects an increase in surface acid density at the external surfaces, eventually resulting in an improvement in ethylene selectivity. The results obtained from density functional theory (DFT) calculations also reveal that ethylene is possibly produced over the Brønsted acid sites located at the external surfaces, whereas diethyl ether (DEE) formation is the predominant pathway over the internal acid sites of MFI. From these findings, hierarchical ZSM-5 nanosheets consisting of a high fraction of active sites at the external silanol surfaces can significantly enhance the selectivity of ethylene from ethanol/bioethanol dehydration.

---

1 Introduction

Ethylene is one of the most crucial light olefinic compounds in the petrochemical industry because it is widely used as a building block for various chemicals ranging from plastics and solvents to antifreeze agents.^1,2 Although there are many vital processes that can be used to produce ethylene, for example, conventional naphtha cracking using steam or thermal techniques,^1,2 there are serious problems with fossil fuel depletion, which has been increasing over the past few decades, and environmental and global warming issues. In order to circumvent these problems, alternative resources are considered as very important feedstocks for chemical production.^3,4 Bioethanol is a well-known renewable organic substance obtained via the fermentation process of biomass.^5–7 Because of a large increase in its production volume compared to its limited demand in a few applications, it is anticipated that excess bioethanol will allow the development of novel technology for the production of value-added compounds from bioethanol in the very near future.^8,9 Therefore, the catalytic conversion of bioethanol via the dehydration route is one of the most fascinating processes for the production of high value-added chemicals such as ethylene.

The dehydration of ethanol is typically performed using various solid-acid catalysts, in particular, metal oxides,^10–15 heteropoly acids,^16–20 and zeolites.^5,21–34 There have been several reports on the catalytic performance of ethanol conversion on acidic zeolites obtained under different reaction conditions.^5,21–34 Compared to Al₂O₃ and amorphous SiO₂, zeolites with various topologies exhibit significantly higher yields of ethylene.^35,36 However, at low temperatures (<573 K) the main product is diethyl ether (DEE).^35,36 Because the dehydration of ethanol to ethylene is an endothermic reaction,^35,36 a high reaction temperature is required (>723 K), and therefore a large proportion of higher hydrocarbons or/and aromatic products is also observed (>35% of the yield of hydrocarbons at 625 K using H-ZSM-5).³⁷ This behavior eventually leads to fast catalyst deactivation due to coke formation inside their microporous structures.

An alternative way to improve the efficiency of zeolite-based catalysts for the selective dehydration of ethanol to ethylene is the use of hierarchical zeolites due to their outstanding properties, such as high external surface area and porosity, resistance to coke formation, and improved catalyst lifetime.³⁸ Typically, such materials have been successfully prepared with one of the following two main approaches: (i) top-down or post treatment methods via desilication (alkaline treatment) or dealumination (acid treatment); (ii) bottom-up or direct synthesis methods via using hard templates (carbon allotropes) or soft templates (e.g., surfactants, cationic polymers, amphiphilic organosilane surfactants).^39–47 Recently, various hierarchical zeolites with nanosheet assembled structures have been thoroughly developed using designed amphiphilic organosilane surfactants or quaternary ammonium/phosphonium cations.^48–54 Such materials exhibit fascinating properties, in particular, the presence of interstitial pore space between nanosheet layers.^48,49,51,54

Although there are some reports on modified HZSM-5 zeolite catalysts for ethanol dehydration at low temperatures (220–350 °C), exhibiting acceptable values for the ethanol conversion (70–100%) and ethylene selectivity (86–100%) (see the ESI Table S1†), the feasibility of using hierarchical zeolite nanosheet assemblies for selective ethanol dehydration to ethylene, and in particular, that of applying them in the dehydration of bioethanol feedstock, has not yet been demonstrated even though they would have many benefits, such as the two-fold advantage of external surface areas and unique acidic surface properties. In addition, an interpretation of the structural and acidic properties of zeolite nanosheets relating to their catalytic activity and product selectivity has not yet been studied.

In the present study, we have attempted to modify the ZSM-5 zeolite with nanosheet structures in a wide range of Si/Al ratios to improve the selective production of ethylene from ethanol and bioethanol resources. The modified nanosheet catalysts have been successfully synthesized via a one-pot hydrothermal synthesis in the presence of a quaternary phosphonium cation (tetrabutylphosphonium hydroxide, TBPOH) as a bifunctional structure-directing agent (SDA). Their structures were characterized by means of X-ray diffraction (XRD), scanning electron microscopy (SEM) and transmission electron microscopy (TEM), NH₃ temperature programmed desorption (NH₃-TPD), pyridine adsorption experiments and ²⁷Al and ²⁹Si MAS NMR spectroscopy. In addition, to maximize the efficiency of the catalysts in terms of activity, the ethylene selectivity and catalyst lifetime for the dehydration of ethanol to ethylene, several experimental parameters, including the Si/Al ratio of the catalyst, the reaction temperature, and the weight hourly space velocity (WHSV), have been systematically studied. Finally, mechanistic perspectives of ethanol dehydration to ethylene and DEE on nanosheet zeolite surfaces were also investigated using a density functional theory (DFT) approach to gain insights into the effect of hierarchical surfaces on the promotion of the catalytic selectivity of ethylene.

2 Experimental section

2.1 Preparation of the catalysts

Hierarchical MFI nanosheets with various Si/Al ratios were synthesized using a modified literature method.^51,54 In a typical procedure, the first aqueous solution consisted of 8.7 g of tetraethyl orthosilicate (Sigma-Aldrich, >99.0%), 8.6 g of tetrabutylphosphonium hydroxide (TBPOH) solution (Sigma-Aldrich, 40 wt% in H₂O) as the SDA, and aluminium isopropoxide (Sigma-Aldrich, >98.0%) in varying amounts to vary the Si/Al ratio of the samples. To prepare the second synthesis solution, 0.0208 g of sodium hydroxide (Sigma-Aldrich, >98.0%) was dissolved in 2.3175 g of deionized water. Subsequently, the first solution was added dropwise into the second solution under vigorous stirring at room temperature for 12 h to obtain a homogeneous mixture. The gel solution was then transferred to a Teflon-lined stainless steel autoclave and crystallized at 130 °C for 48 h. The solid powder was precipitated using a centrifugal technique and washed with deionized water until neutral pH of the supernatant fraction was obtained. Finally, the sample was dried overnight at 110 °C and then calcined at 650 °C for 8 h to remove the SDA and the other contaminants. The synthesized solid was converted to H⁺ form with three consecutive ion-exchange steps with 0.1 M NH₄Cl solution at 80 °C for 2 h. The obtained powder was dried and finally calcined at 650 °C for 6 h. The catalyst was defined ZSM5-HieNS(x), where x is the actual Si/Al ratio.

To prepare the conventional ZSM-5 catalyst, a synthesized precursor containing 3.50 g of tetraethyl orthosilicate (Sigma-Aldrich, >99.0%), 2.02 g of tetrapropylammonium hydroxide solution (Sigma-Aldrich, 1 M in H₂O), 0.0276 g of sodium aluminate (Sigma-Aldrich, >99.5%, anhydrous) and 0.0693 g of sodium hydroxide (Sigma-Aldrich, >98.0%) in 10.465 g of deionized water was stirred at room temperature for 2 h. Then, this was transferred to a stainless-steel autoclave for hydrothermal synthesis at 180 °C for 72 h.^51,55 Finally, a solid sample was calcined at 650 °C for 8 h. The acid catalyst was obtained by applying the above-mentioned ion-exchange procedure. The obtained conventional ZSM-5 was denoted ZSM5-CON(x), where x is the actual Si/Al ratio.

2.2 Characterization of the catalysts

Powder X-ray diffraction (XRD) patterns were obtained using a Bruker D8 ADVANCE diffractometer and CuK_α radiation (30 kV, 40 mA) with a step size of 0.02° and a scan rate of 1° min⁻¹ in the two theta (2θ) range of 5 to 50°. Scanning electron microscopy (SEM) and transmission electron microscopy (TEM) were used to reveal the catalyst morphologies and were performed using JEOL JSM-7600F (at 2 kV) and TECNAI 20 TWIN (at 200 kV), respectively.

Wavelength-dispersive X-ray fluorescence spectrometry (WDXRF) was used to analyze the elemental composition of the samples and was performed using a Bruker S8 Tiger instrument. The textural properties were determined through an N₂ adsorption/desorption technique at 77 K performed on BELSORP-max apparatus. The BET specific surface area (S_BET), and the external surface area (S_ext), together with the micropore volume (V_micro), and the mesopore size distribution were calculated using the Brunauer–Emmett–Teller (BET) method, the t-plot method, and the Barrett–Joyner–Halenda (BJH) model, respectively.

To investigate the acidic properties, NH₃ temperature programmed desorption (NH₃-TPD) profiles were recorded using a BELL-CATII analyzer. Typically, a catalyst (0.05 g) was pretreated under the flow of He at 500 °C with a heating rate of 10 °C min⁻¹. Subsequently, a diluted ammonia feed (7.5 v/v% of NH₃ in He) was introduced for 0.5 h. The NH₃-TPD profiles were recorded by increasing the temperature from 100 to 700 °C with a heating rate of 10 °C min⁻¹ in the flow of He. To investigate the coke content, O₂ temperature programmed oxidation (O₂-TPO) was performed on a BELL-CATII analyzer with the same pretreatment conditions as those in the NH₃-TPD experiments. After pretreatment, the feed of a diluted oxygen (5 v/v% of O₂ in He) with a total mass flow rate of 50 mL min⁻¹ was introduced, and the temperature was linearly increased from 50 to 700 °C using a heating rate of 5 °C min⁻¹. FTIR spectra were recorded in the range 4000–800 cm⁻¹ using a Bruker Vertex V70v instrument. The spectra were gained at a 2 cm⁻¹ resolution and averaged over 64 scans. Firstly, the zeolite sample was heated to 550 °C at a rate of 2 °C min⁻¹ under 20 v/v% of O₂ in He. To determine the density of acid sites in the zeolites, pyridine was introduced for about 10 min into the chamber at its vapor pressure at room temperature. Then, the pyridine molecules were desorbed by evacuation for 1 h at 150 °C. Afterwards, a spectrum was recorded at 150 °C. To calculate the total amount of Brønsted and Lewis acid sites, the molar extinction coefficient values 0.73 and 1.11 cm μmol⁻¹ were applied, respectively.

The nature of the aluminium sites and the environment of the silicon species in the zeolite framework were investigated using ²⁷Al and ²⁹Si MAS NMR spectroscopy operating at the magnetic fields 17.6 T and 9.4 T, respectively, performed on an AVANCE III HD (400 MHz) Digital NMR spectrometer (Bruker).

2.3 Catalytic studies of ethanol and bioethanol dehydration

To carry out the catalytic study of ethanol/bioethanol dehydration, the reaction was performed using a continuous fixed-bed tubular reactor with a tube with an inner diameter of 3/8 inches. The catalyst was sieved into particle sizes in the range 450–850 μm to control the particle size distribution and packed in the middle of the reactor tube between thin layers of glass wool. Prior to the catalytic test, the catalyst was pretreated at 500 °C under the flow of N₂ (50 mL min⁻¹) for 2 h. Subsequently, the denatured ethanol (99.99%, QRec) or bioethanol (Sapthip company) in N₂ carrier gas (1 : 2.5 molar ratio) was introduced into the reactor at the reaction temperature range 200 to 400 °C with weight hourly space velocities (WHSV) in the range 10.0 to 30.0 h⁻¹. To prevent the condensation of liquid products, the inlet and outlet zones were heated to 185 °C. The products were analyzed using an online gas chromatograph (GC) (Agilent 7820A) equipped with an FID detector and a PoraBOND Q capillary column (10 m × 3 μm × 0.25 mm) at an interval time of 1 h. The chromatogram of the reaction mixture is shown in Fig. S1.† The mass balance of all the experiments was calculated from the calibration curve of each component and their values were in the range 99.03 ± 0.07% (Fig. S2†). The conversion of ethanol (X_EtOH) and the product selectivity (S_i) were estimated using eqn (1) and (2), respectively:where (n_EtOH)_i, (n_EtOH)_t, n_i and ∑n_j are the initial number of moles of ethanol, the number of moles of ethanol at a certain time, the number of moles of the desired product i, and the total number of moles of all the products j, respectively.

2.4 Computational studies

Density functional theory (DFT) calculations were performed to investigate the mechanism of the dehydration of ethanol to ethylene over the Brønsted acid sites on different surfaces of the MFI framework. The 12T quantum cluster models taken from the MFI zeolite lattice structure⁵⁶ were used to represent the Brønsted acid sites at the external surfaces of hierarchical ZSM-5 and the internal surfaces of the bulk MFI framework as shown in Fig. 9A and B, respectively. To keep the crystalline structure of the MFI framework, the dangling bonds of the surface oxygen atoms were terminated by H atoms at the distance 1.47 Å ^57–59 and the Si–H bonds were aligned in the directions of the Si–O bonds of the zeolite structure. To represent the active site at the external MFI surfaces, the 12T model was generated from the outer surface of the zeolite⁶⁰ (Fig. 9A). An aluminium atom was substituted at the T12 position and a Brønsted acid site was generated by adding a hydroxyl group on this aluminum site.⁶¹ In addition, three hydroxyl groups were added to the silicon atoms at the T7, T9, and T10 positions to represent surface silanol groups, as can be seen in Fig. 9A.^60,61 For the model to represent the active site at the internal surface of the MFI framework corresponding to the bulk MFI catalyst, it was constructed using the 12T quantum cluster generated which was composed from the intersecting straight and zigzag channels of MFI with the replacement of an aluminium atom at the T12 position to generate the Brønsted acid site⁵⁸ (Fig. 9B).

The M06-2X density functional was used to perform all calculations. This method can describe the van der Waals interactions and is suitable for studying various catalytic reactions over zeolite surfaces.^57,62,63 During the geometry optimizations, the 12T active region and adsorbates were allowed to relax, while the remaining hydrogen atoms were fixed at crystallographic coordinates. All calculations were performed with the GAUSSIAN 09 code using the basis set 6-31G(d,p). The interaction energies (ΔE) of the intermediates and products were calculated with eqn (3):

ΔE = E_complex − (E_zeolite + E_adsorbate) (3)

where ΔE is the interaction energy of the system (kcal mol⁻¹), E_complex is the total energy of the complex structure of the zeolite and adsorbate molecules (kcal mol⁻¹), E_zeolite is the total energy of the isolated zeolite surfaces (kcal mol⁻¹), and E_adsorbate is the total energy of the isolated adsorbate (kcal mol⁻¹).

3 Results

3.1 Physicochemical properties of the hierarchical ZSM-5 catalysts (ZSM5-HieNS)

The XRD patterns of all the synthesized samples indicate that the synthesized samples contain the MFI framework, for which the main characteristic peaks appeared at 2θ of 7.8, 8.7, 23.1, 23.8 and 24.3°, corresponding to (101), (111), (501), (303) and (313),^64,65 respectively, as shown in Fig. 1A. The relative crystallinity calculated with respect to a conventional MFI (MFI-CON(50)) as a reference, and the average crystallite size calculated with the Scherrer equation from the peak area in the range of 2θ 22 to 25° ^55,66 are summarized in Table S2.† Clearly the relative crystallinity values of the hierarchical nanosheets are above 90% compared to those of the bulk zeolite, implying that the crystalline structures are well-preserved even in the presence of hierarchical structures. Interestingly, an increase in the relative crystallinity relating to a slightly larger average crystallite size is observed as a function of increasing Si/Al ratio.

Fig. 1 (A) XRD patterns of (a) MFI-CON(38), (b) MFI-HieNS(37), (c) MFI-HieNS(75), (d) MFI-HieNS(103), and (e) MFI-HieNS(159); (B) SEM and TEM images of (a1 and b1) MFI-CON(38), and (a2 and b2) MFI-HieNS(37).

The morphologies of the as-synthesized samples were observed using SEM and TEM microscopic techniques (Fig. 1B). As expected, MFI-CON(38) exhibits a bulk structure with cubic-shaped crystals (Fig. 1B(a1 and b1)). In contrast, the MFI-HieNS samples exhibit a different structure due to spherical assemblies of nanolayers (Fig. 1B (a2 and b2), and Fig. S3† for the hierarchical samples with different Si/Al ratios). In addition, the particle size distribution and nanosheet thickness of the nanolayers are demonstrated in Fig. S4 and S5,† respectively. Compared to MFI-CON(38), MFI-HieNS(37) has a smaller particle size which is in the range 224.09 ± 41.25 nm. However, when increasing the Si/Al ratio, the particle size is slightly increased to 255.87 ± 31.48 nm for the MFI-HieNS(159) sample. The thickness of the as-synthesized nanosheets is approximately 8–9 nm.

The textural properties of the as-synthesized samples were investigated with N₂ physisorption measurements as shown in Fig. 2 and Table 1. The adsorption–desorption isotherm of MFI-CON(38) corresponds to a type I isotherm due to micropore filling in the microporous network at a low relative pressure (P/P₀) (Fig. 2A(a)). However, the nanosheet series of MFI catalysts exhibit a combined isotherm which is between type I and IV isotherms, in which a hysteresis loop at high P/P₀, due to a capillary condensation within the additional secondary pores, is observed (Fig. 2A). This makes it clear that the zeolite nanosheets are composed of microporous features due to their zeolite characteristics together with mesopores/macropores obtained from the assemblies of nanolayers and the inter-particle voids. This behavior eventually leads to an increase in external surface area and pore volume (Table 1).

Fig. 2 (A) N₂-sorption isotherms and (B) NH₃-TPD profiles of (a) MFI-CON(38), (b) MFI-HieNS(37), (c) MFI-HieNS(75), (d) MFI-HieNS(103), and (e) MFI-HieNS(159).

Table 1 Textural properties of all the synthesized samples obtained with N₂ physisorption measurements

Sample	Si/Al^a	Surface area (m^2 g^−1)			Pore volume (cm^3 g^−1)			S ext/SBET^h	V ext/Vtotal^i
		S BET	S Micro	S ext	V total	V micro	V ext
a Si/Al ratio was determined using an XRF technique. b S BET: BET specific surface area. c S Micro: micropore surface area. d S ext: external surface area. e V total: total pore volume obtained at P/P0 = 0.99. f V micro: micopore volume. g V ext: external pore volume calculated from the difference between the total pore volume and micropore volume. h The fraction of external surface area. i The proportional external volume.
MFI-CON(38)	38	405	390	15	0.22	0.18	0.03	0.04	0.16
MFI-HieNS(37)	37	451	286	165	1.02	0.11	0.90	0.37	0.89
MFI-HieNS(75)	75	562	412	150	1.10	0.22	0.88	0.27	0.80
MFI-HieNS(103)	103	505	352	153	1.07	0.17	0.89	0.30	0.84
MFI-HieNS(159)	159	589	433	156	1.05	0.25	0.81	0.26	0.77

---

The chemisorption profiles and acidic properties of all the samples were investigated using the NH₃ temperature programmed desorption (TPD) technique as shown in Fig. 2B and Table 2. All the synthesized samples exhibit two predominant NH₃ desorption bands at the low-temperature desorption peak (190 to 205 °C) and the high-temperature desorption peak (375 to 400 °C) corresponding to the characteristics of weak acid sites and strong acid sites, respectively.⁶⁷ Typically, the weak acid sites relate to the non-framework Al species or the Si–OH in the structural defects, which can interact with base molecules.⁶⁸ As can be seen in Table 2, MFI-HieNS(37) and MFI-CON(38) have an equivalent total acid density, however the ratio of the weak to the strong acid density of the nanosheet samples obviously increases. This might be explained by the change in the microporous structure and the addition of secondary mesoporous networks leading to an increase in the external surface properties. Obviously, the ratio of the weak to the strong acid densities continuously decreases as a function of the Si/Al ratio, relating to a decrease in the number of strong acid sites when increasing the Si/Al ratio.

Table 2 Acidic properties and the proportion of silanol defects of the catalysts obtained from the NH₃-TPD and ²⁹Si MAS NMR techniques

Sample	Desorption temperature (°C)		Acidity amount (μmol g^−1)			W/S^f	Q2/Q4^g	Q3/Q4^h
	Low DES^a	High DES^b	Weak^c (W)	Strong^d (S)	Total^e
a Low temperature desorption peak (DES). b High temperature desorption peak (DES). c Weak acid density. d Strong acid density. e Total acid density. f The ratio of the weak and strong acid density; the weak acid density, strong acid density and total acid density were calculated using a deconvoluted Gaussian distribution. g The ratio of Q2 to Q4. h The ratio of Q3 to Q4 obtained using the ^29Si NMR technique.
MFI-CON(38)	201	396	90	130	220	0.69	0.50	4.83
MFI-HieNS(37)	203	376	100	110	221	0.91	0.74	7.09
MFI-HieNS(75)	202	385	70	100	170	0.70	0.70	6.81
MFI-HieNS(103)	196	379	30	60	90	0.50	0.63	4.91
MFI-HieNS(159)	193	374	20	60	80	0.33	0.59	3.89

---

To investigate the nature of the aluminium and silicon species in the zeolite framework, the ²⁷Al and ²⁹Si MAS-NMR techniques were used, respectively. The ²⁷Al-NMR spectra of all the synthesized samples are illustrated in Fig. S6.† The tetrahedral aluminium in the frameworks corresponds to the peak at the chemical shift 55 ppm, while the extra-aluminum framework (EFAl), an oligomeric alumina with a low degree of hydration/hydroxylation, relates to the signal at the chemical shift 0 ppm.⁶⁹ Although hierarchical ZSM-5 exhibits the presence of EFAl species, a large proportion of the aluminum sites are due to the tetrahedral aluminium framework. This is also consistent with the pyridine adsorption results, confirming the presence of both Brønsted acid sites and Lewis acid sites (in Fig. S7 and Table S3†). Interestingly, the content ratio of the Brønsted acid sites to the Lewis acid sites of MFI-HieNS is obviously smaller than that of MFI-CON. The presence of EFAl species might be related to the removal of aluminium during the calcination at high temperature. The presence of EFAl species together with that of the Brønsted acid sites (BAS) of tetrahedral aluminium in the framework also provides a synergistic effect between cationic EFAl and vicinal BAS and can promote some catalytic reactions, such as catalytic cracking and dehydration.^70–72

To reveal the environment of the Si and Al frameworks, ²⁹Si-NMR spectra were obtained and are illustrated in Fig. 3. They indicate the presence of the following three characteristic peaks: (i) Si surrounded by two Si or Al atoms and two hydroxyl groups (Q₂); (ii) Si surrounded by three Si or Al atoms and one hydroxyl group (Q₃); (iii) Si surrounded by four Si or Al atoms (Q₄). They appear at the chemical shifts −93, −103, and −113 ppm for Q₂, Q₃, and Q₄, respectively.⁷³ Compared to the conventional HZSM-5 (MFI-CON(38)), the hierarchical nanosheet with a similar Si/Al ratio (MFI-HieNS(37)) exhibits a significantly higher ratio of Q₂/Q₄ and Q₃/Q₄ (Table 2), which corresponds to a large number of Si or Al atoms sitting on the external surfaces or zeolite defects to form silanol surfaces. It is therefore reasonable to assume that the modified zeolites are composed of a large number of silanol defect sites or disturbed tetrahedral coordinated aluminium sites due to increased external surface area obtained by introducing a secondary mesoporous network. In addition, the Q₂/Q₄ and Q₃/Q₄ ratios obviously increase with decreasing Si/Al ratio. The reason for this increase in the ratio of silanol defects to tetrahedral frameworks relates to the fact that a zeolite with a low Si/Al ratio demonstrates a high possibility of the formation of a disturbed aluminum tetrahedral coordinated state of aluminum,⁷⁴ which is also exhibited by the broad signal in the ²⁷Al NMR spectra in the range 25 to 50 ppm (Fig. S6B†). In addition, the high ratio of silanol surfaces to bridging hydroxyls (Brønsted sites) of the hierarchical zeolite nanosheet can be confirmed by the high proportion of O–H stretching from terminal Si–OH groups at 3740 cm⁻¹ with respect to bridging hydroxyls (Brønsted sites), which appeared at 3600 cm⁻¹, whereas this ratio is obviously lower in the case of the conventional zeolite (Fig. S8†).

Fig. 3 ²⁹Si-NMR spectra of (A) MFI-CON(38), and (B–E) MFI-HieNS with the Si/Al ratios 37, 75, 103, and 159, respectively.

3.2 Catalytic conversion of ethanol through dehydration to ethylene

3.2.1 Effect of the reaction parameters. The influence of the reaction temperature on the catalytic performance was investigated at low to moderate temperatures (200–350 °C) under atmospheric pressure. As can be seen in Fig. 4A, the conversion of ethanol, as expected, increases dramatically as a function of reaction temperature. At a low temperature (200 °C) the main product corresponds to diethyl-ether (DEE), whereas a very low selectivity of ethylene is observed (3.2%). In contrast, when increasing the reaction temperature to above 300 °C the selectivity of ethylene is high and approaches 98.0% at a high ethanol conversion level of 96.5%. This is consistent with the view that the pathway to produce DEE from ethanol is an exothermic reaction as shown in eqn (4), whereas ethanol dehydration to ethylene is an endothermic reaction (eqn (5)), which requires high temperatures.^35,75

2C₂H₅OH → C₂H₅OC₂H₅ + H₂O, ΔH⁰₂₉₈ = −25.1 kJ mol⁻¹ (4)

C₂H₅OH → C₂H₄ + H₂O, ΔH⁰₂₉₈ = +44.9 kJ mol⁻¹ (5)

Fig. 4 (A) Ethanol conversion as a function of time-on-stream (TOS) and (B) conversion and selectivity of the products obtained at 15 h of TOS with different reaction temperatures: (a) 200 °C, (b) 250 °C, (c) 300 °C, (d) and 350 °C at a WHSV of 10 h⁻¹, and the molar ratio of N₂ to EtOH 2.5 : 1 over the MFI-HieNS(37) catalyst.

The effect of space velocity on the catalytic performance was studied by variation in terms of the weight hourly space velocity, WHSV (10, 17, 23, and 30 h⁻¹), under atmospheric pressure at the reaction temperature 300 °C over the MFI-HieNS(37) catalyst. As expected, the activity of ethanol conversion significantly decreases from 91.9% to 53.6% when increasing the WHSV from 10 to 30 h⁻¹ (Fig. 5A). The selectivity of ethylene decreases from 85.7% to 20.7%, whereas the selectivity of DEE increases from 14.3% to 79.2% at WHSV values of 10 and 30 h⁻¹, respectively, as shown in Fig. 5B. The reason for the change in activity and product distribution as a function of WHSV relates to the fact that a low space velocity allows a long time for the reaction mixture to be in contact with the catalyst, and therefore a high conversion of the reactant can be observed.^76,77 In addition, the higher ethylene selectivity at a low WHSV might be explained by the direct decomposition of diethyl ether (DEE) to ethylene. This reaction pathway is the competing reaction with direct dehydration of ethanol to ethylene, and may occur depending on the relative partial pressures of ethanol and DEE, which can be controlled with specific reaction conditions.^35,78,79 Therefore, it is reasonable to propose that the reaction pathway when using a low space velocity of reactant proceeds via a parallel series reaction, including the following three reactions: (i) direct conversion of ethanol into ethylene; (ii) ethanol transformation to DEE; (iii) decomposition of DEE to ethylene⁷⁹ (eqn (6)).

C₂H₅OC₂H₅ → 2C₂H₄ + H₂O, ΔH⁰₂₉₈ = +115.7 kJ mol⁻¹ (6)

Fig. 5 (A) Ethanol conversion as a function of time-on-stream (TOS) and (B) conversion and selectivity of the products obtained at 15 h of TOS with different WHSV values: (a) 30 h⁻¹, (b) 23 h⁻¹, (c) 17 h⁻¹, and (d) 10 h⁻¹ at the reaction temperature 300 °C and the molar ratio of N₂ to EtOH 2.5 : 1 over the MFI-HieNS(37) catalyst.

3.2.2 Effect of the hierarchical structure and Si/Al ratio. To illustrate the beneficial effect of the hierarchical nanosheet structures on the catalytic reaction of ethanol dehydration, a bulk catalyst with a similar Si/Al ratio (MFI-CON(38)) was used as the reference catalyst (Fig. 6). Although a high activity can be observed on the fresh catalyst, ethanol conversion gradually decreases as a function of TOS. Clearly, the ethanol conversion reduces from 89.3% to 59.4% even after a few hours of TOS. In strong contrast to this, the high conversion over the MFI-HieNS(37) is still observed (∼91.9) as shown in Fig. 6. Compared to that of the (MFI-CON(38)) catalyst, not only has the activity been changed over the MFI-HieNS(37) sample, but also the product selectivity significantly differs between the two. Interestingly, a high selectivity of ethylene (almost 90%) can be observed over MFI-HieNS(37). However, the main product over (MFI-CON(38)) is DEE with a selectivity of 60.4%, whereas only 39.6% for the ethylene selectivity is observed.

Fig. 6 (A) Ethanol conversion as a function of time-on-stream (TOS) and (B) conversion and selectivity of the products obtained at 15 h of TOS using (a) MFI-CON(38) and MFI-HieNS with various Si/Al ratios: (b) MFI-HieNS(37), (c) MFI-HieNS(75), (d) MFI-HieNS(103), and (e) MFI-HieNS(159) at the reaction temperature 300 °C, a WHSV of 10 h⁻¹, and the molar ratio of N₂ to EtOH 2.5 : 1.

In addition, the catalytic performance of MFI-HieNS with different Si/Al ratios was also investigated (Fig. 6B). This clearly shows that a high ethanol conversion (∼86.9% to 91.9%) is still observed irrespective of the Si/Al ratio. However, the product selectivity significantly changes. For example, the ethylene selectivity gradually decreases when increasing the Si/Al ratio. The highest selectivity of ethylene ∼85.5% is found when using the MFI-HieNS(37) catalyst. The reason for this increase in the ethylene selectivity as a function of decreasing Si/Al ratio relates to the fact that a lower Si/Al ratio corresponds to a high acid density and this may increase the possibility of primary dehydration of the alcohol and decomposition of the ether intermediates to light olefins.^38,80,81 It is well-known that the major disadvantage of a bulk zeolite for ethanol conversion is the short catalyst lifetime. To further confirm the benefits of hierarchical structures for catalytic stability, a stability test was carried out over the synthesized catalysts under the same reaction conditions, as shown in Fig. 7A. Clearly, MFI-CON(38) shows fast deactivation of the catalyst even after 5 h of TOS and the time taken to reach 50% ethanol conversion is less than 20 h. Compared to conventional HZSM-5 with a similar Si/Al ratio (MFI-CON(38)), MFI-HieNS(37) has greatly improved catalytic stability. This makes it clear that the hierarchical structures can enhance the stability of the catalyst. In addition, the stability of the catalyst also depends on the Si/Al ratio. It was found that a high Si/Al ratio (MFI-HieNS(103)) can increase the catalyst stability. However, the activity of a sample in which the Si/Al ratio is too high (MFI-HieNS(159)) becomes lower. The reason for this decrease in activity relates to the fact that there are insufficient acid sites, resulting in suppression of the catalytic reaction.

Fig. 7 (A) Ethanol conversion as a function of time-on-stream (TOS) and (B) total coke content measured using the O₂-TPO technique over MFI-HieNS with various Si/Al ratios: (a) MFI-CON(38), (b) MFI-HieNS(37), (c) MFI-HieNS(103), and (d) MFI-HieNS(159) at the reaction temperature 300 °C, a WHSV of 10 h⁻¹, and the molar ratio of N₂ to EtOH 2.5 : 1.

To verify the accumulated coke deposit in the zeolite structures, the coke contents derived from O₂-TPO profiles (Fig. S9†) to represent the coke species over MFI-HieNS(37) and MFI-CON(38) are shown in Fig. 7B. The results indicate that MFI-HieNS(37) can reduce the amount of the coke species (0.127 mmol_coke g_cat⁻¹), whereas a significantly higher amount of coke can be observed over MFI-CON(38) (0.151 mmol_coke g_cat⁻¹). In addition, when increasing the Si/Al ratio, the total amount of coke deposits is obviously reduced as can be seen in Fig. 7 because a low acid density eventually leads to suppression of the hydrocarbon pool pathway.^82,83

To investigate the regeneration of the catalysts, the catalytic performance of the hierarchical MFI catalysts was investigated after catalyst regeneration by the burning of coke species. It was found that hierarchical ZSM-5 exhibits high ethanol conversion and high selectivity of ethylene (∼90–95%) even after regeneration of the catalysts (Fig. S10†). In addition the morphologies of the catalysts can be well preserved (Fig. S11–S12, and Table S4†).

3.2.3 Catalytic conversion of bioethanol through dehydration to ethylene. To extend the scope of ethanol dehydration to the potential application of real bioethanol conversion, we also carried out additional experiments using bioethanol feedstock obtained by the fermentation of sugar cane and cassava chips. In this experiment, the catalytic performance of bioethanol dehydration was studied over MFI-CON(38) and MFI-HieNS(37), which have the same Si/Al ratio and total acid density. The composition of bioethanol was analyzed using GC-MS and is summarized in Table S5.† It was found that the major component was ethanol (∼98.3%) with some impurities, such as 2-methyl-1-propanol (∼0.6%), isopropyl alcohol (∼0.4%), and toluene (∼0.2%). Interestingly, MFI-HieNS(37) exhibits higher activity without a significant decrease in ethanol conversion even after 15 h with a high conversion level of ∼80% (Fig. 8A). However, MFI-CON(38) exhibits fast deactivation of the catalyst after a few hours of TOS and the conversion of bioethanol decreases to 33.7% after 10 h of TOS. The analysis of the product distribution again confirms that MFI-HieNS(37) can selectively produce ethylene with a selectivity of ∼70%, whereas the main product over MFI-CON(38) is DEE with a low selectivity for ethylene (less than 40%). From these observations, it is clear that hierarchical HZSM-5 can greatly enhance the catalytic performance in terms of the activity, catalyst stability, and ethylene selectivity even when applying it to a real commercial bioethanol. This opens up interesting perspectives on the development of highly selective catalysts for ethylene production from commercial bioethanol.

Fig. 8 (A) Bioethanol conversion as a function of time-on-stream (TOS) and (B) selectivity of the products obtained at 15 h of TOS using real bioethanol as the feedstock over MFI-CON(38) and MFI-HieNS(37) at the reaction temperature 300 °C, a WHSV of 10 h⁻¹, and the molar ratio of N₂ to EtOH 2.5 : 1.

3.3 Mechanistic studies

As stated above, the hierarchical ZSM-5 nanosheets dramatically increase the ethylene yield of the ethanol conversion compared to that of the bulk ZSM-5. To understand the reaction mechanism of ethanol dehydration to ethylene, we also carried out a series of calculations using computational approaches. In this study, we focused on the above-mentioned mechanisms, including direct conversion of ethanol to ethylene and ethanol transformation to DEE.⁷⁹

To compare the reaction mechanisms over different structures of catalysts, two different models, composed of the Brønsted acid sites on the external silanol surfaces (containing Q₃ species) and the internal surfaces of the bulk MFI framework (containing Q₄ species), have been used to represent the structures of hierarchical ZSM-5 (Fig. 9A) and a conventional one (Fig. 9B), respectively.

Fig. 9 Optimized structures representing (A) the Brønsted acid site at the external silanol surfaces of hierarchical ZSM-5, and (B) the Brønsted acid site at the internal surfaces of a conventional MFI framework.

To represent the hierarchical ZSM-5 surfaces, in which the external silanol surface (Fig. 9A) is predominant, a hydroxyl group was added to the aluminium atom (T12 site) of the MFI framework to create a Brønsted acid site and other three hydroxyl groups were introduced at the silicon atom of the external surface of MFI.^60,61 As can be seen in Fig. 9A, the bond distances of O4⋯H1s and O4⋯Al are 0.97 and 1.72 Å, respectively. Furthermore, the average bond distance between the Al atom and its three neighboring oxygens (O1, O2, and O3) atoms is 1.77 Å. As for the structure representing the acid site at the internal surfaces of a conventional MFI framework (Fig. 9B), the 12T cluster covers the intersecting straight and zigzag channels, in which an aluminum atom was substituted at the T12 position to generate a Brønsted acid site.⁵⁸ The bond distances of O1⋯Al and O2⋯Al are 1.83 and 1.68 Å, respectively. The bond distance of the Brønsted acid site (O1–H1s) is 0.97 Å.

3.3.1 Reaction mechanisms of ethanol dehydration to ethylene and DEE over ZSM-5. From experimental observations, the mechanism of ethanol dehydration to ethylene has been proposed, including direct conversion of ethanol to ethylene and ethanol transformation to DEE.⁷⁹ To verify the reactivity of both the above-mentioned models representing the Brønsted acid site at the external surfaces of a hierarchical zeolite (Fig. 9A), which also exists due to the bridging Si–(OH)–Al groups at the pore mouth,⁸⁴ and the acid site at the internal surfaces of the conventional zeolite framework (Fig. 9B), mechanistic pathways have been demonstrated as shown in Fig. 10 and the elementary reaction steps are described in eqn (7)–(9) in Scheme 1. In the first step of the reaction, the ethanol molecule is dehydrated over the Brønsted acid site of ZSM-5 to produce an ethoxide intermediate (I) via an ethyl carbenium transition state. Subsequently, the reaction mechanism can be classified into two routes: the direct decomposition of an ethoxide intermediate to ethylene (I → II → III → IV) and the transformation of a second ethanol on the ethoxide surfaces to DEE (I → V → VI → VII → VIII).^85,86

Fig. 10 Proposed ethanol dehydration pathways for ethanol dehydration to ethylene and DEE.

Scheme 1 The elementary reaction steps for dehydration of ethanol to ethylene and diethyl ether over HZSM5.

3.3.2 Influence of the surface structure on the mechanisms of ethanol dehydration to ethylene and DEE. The optimized structures and the energy profile of ethanol conversion to ethylene and DEE over the Brønsted acid sites located at the external surfaces of hierarchical ZSM-5 and the internal surfaces of a bulk MFI framework are shown in Fig. 9, S13, S14† and 11, respectively. This reaction mechanism is initiated from the reaction of the ethoxide intermediates, as shown in Fig. 11A(I) and (I′). As for the direct decomposition of the ethoxide species to ethylene, the hydrogen of the ethoxy is abstracted by the oxygen framework of the zeolite (Fig. 11A(II) and (II′)). Over the external surface model, the bond length of C1⋯H1e of the ethoxy species is elongated from 1.09 to 1.21 Å, and then the H1e atom moves closer to O2 of the zeolite framework with a bond length of 1.62 Å. This step requires an activation energy of 41.2 kcal mol⁻¹. As for the internal surface model of MFI, a similar process can be observed and this step requires an activation energy of 36.7 kcal mol⁻¹. Finally, the ethylene product is formed over the external and internal surfaces of MFI (Fig. 11A(III) and (III′)).

Fig. 11 Energy profiles of (A) ethanol dehydration to ethylene (I → II → III → IV) and (B) ethanol conversion to diethylether (DEE, I → V → VI → VII → VIII) over the Brønsted acid sites at the external surfaces of hierarchical ZSM-5 (black solid-line), and the internal surfaces of the conventional ZSM-5 framework (blue dash-line).

In strong contrast to this, ethanol dehydration to DEE starts from the co-adsorption of a second ethanol over the ethoxide surface (Fig. 11B(V) and (V′)). The oxygen atom of the co-adsorbed ethanol moves closer to the carbon atom of the ethoxide species, and the hydrogen atom of the hydroxy group of the second ethanol is simultaneously transferred to the oxygen atom of the zeolite framework. As for the external surface model of MFI, the bond distance between the O5 of ethanol and the C1 of the ethoxy species is shortened from 3.10 to 2.47 Å. Meanwhile the bond distance between H1e and O5 of the hydroxy group of ethanol is elongated from 0.97 to 0.98 Å and the H1e of the hydroxy group interacts with the O2 of the zeolite framework with a bond length of 1.76 Å. This reaction step requires an activation energy of 45.5 kcal mol⁻¹. Compared to that in the external surface model, the transition state over the internal surfaces of MFI can occur by a similar pathway, however it requires a significantly lower activation energy of 25.6 kcal mol⁻¹. Finally, diethyl ether is formed over the external and internal surfaces of MFI (Fig. 11B(VII) and (VII′)). As expected, ethoxide decomposition to ethylene and DEE requires lower activation energies over the internal surfaces than over the external surfaces of the MFI zeolite. This behavior relates to the higher acid strength of the Brønsted acid sites at the internal surfaces compared to that of the acid sites at the external surfaces.⁸⁷

From the calculation observations, the ethoxide species can be converted to either ethylene or DEE as products over the external surfaces of hierarchical ZSM-5 and the internal surfaces of the bulk MFI framework. However, ethylene formation possibly occurs over the Brønsted acid site at the external surfaces because the surface silanol groups can induce a charge between the intermediate molecules and the zeolite surfaces. In contrast, diethyl ether formation is preferably carried out over the internal acid sites, confirmed by a large difference in activation energies between ethylene formation and the DEE pathway.

These observations are also consistent with the experimental point of view that ethylene is possibly produced over the hierarchical nanosheets, which have the characteristic of the external silanol surfaces (Q2 and Q3 species), whereas diethyl ether (DEE) formation is the predominant pathway over the internal Brønsted acid sites of the bulk MFI. It is therefore reasonable to assume that the hierarchical ZSM-5 nanosheets consisting of a high fraction of external surfaces compared to bulk ZSM-5 can significantly enhance the selectivity of ethylene from ethanol dehydration. These observations also support the details obtained from previous works which reveal that ethylene is possibly generated from direct ethanol conversion or indirect generation from ether, or that both routes contribute to ethene formation.^88,89 In particular, the surface acid sites of the hierarchical zeolite can promote direct ethanol conversion to ethylene.

4 Conclusions

Hierarchical MFI zeolite nanosheet catalysts have been successfully synthesized via a one-pot hydrothermal method with the aid of TBPOH as a bifunctional SDA. These zeolites had improved physicochemical properties with outstanding external surface area and appropriate acid properties due to a large proportion of external surfaces in the hierarchical structure. To illustrate the beneficial effect of zeolite nanosheets towards green and sustainable catalytic conversion of renewable resources to high value-added chemicals, direct bioethanol dehydration to ethylene over Brønsted acid MFI nanosheets has been studied from both experimental and theoretical points of view. With their superior properties, the hierarchical ZSM-5 nanosheets can greatly enhance ethanol conversion (>90%), with a high selectivity of ethylene (>85%) and high catalyst stability in the ethanol dehydration process at moderate temperatures (300–350 °C). The computational point of view reveals that ethylene is possibly produced over the Brønsted acid sites at the external surfaces whereas diethyl ether formation is the predominant pathway over the internal Brønsted acid sites of MFI. This information provides insight into the fact that the Brønsted acid sites at the external surfaces of the hierarchical ZSM-5 nanosheets can significantly enhance the selectivity of ethylene from ethanol dehydration. This opens up new perspectives on the hierarchical structures not only improving the catalytic performance due to a reduction in coke formation but also the weak acid sites at the external surfaces enhancing the selectivity of ethylene.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by grants from the Vidyasirimedhi Institute of Science and Technology (VISTEC), the Thailand Research Fund (TRF) (MRG6180099), the Office of Higher Education Commission (OHEC), and the Junior Research Fellowship Program of the French Embassy in Thailand. This work has also been partially supported by the National Nanotechnology Center (NANOTEC), NSTDA, Ministry of Science and Technology, Thailand, through its program of Research Network NANOTEC (RNN). Furthermore, the authors would like to acknowledge the Frontier Research Center (FRC) at VISTEC for the technical support.

References

1. Y. Gucbilmez, T. Dogu and S. Balci, Ind. Eng. Chem. Res., 2006, 45, 3496–3502.
2. C. J. Pereira, Science, 1999, 285, 670–671.
3. B. Liu, L. France, C. Wu, Z. Jiang, V. L. Kuznetsov, H. A. Al-Megren, M. Al-Kinany, S. A. Aldrees, T. Xiao and P. P. Edwards, Chem. Sci., 2015, 6, 5152–5163.
4. Y. Li, L. Li and J. Yu, Chem, 2017, 3, 928–949.
5. J. Bi, X. Guo, M. Liu and X. Wang, Catal. Today, 2010, 149, 143–147.
6. Q. Sheng, K. Ling, Z. Li and L. Zhao, Fuel Process. Technol., 2013, 110, 73–78.
7. L. Wang, R. Templer and R. J. Murphy, Energy Environ. Sci., 2012, 5, 8281–8293.
8. J. Sun and Y. Wang, ACS Catal., 2014, 4, 1078–1090.
9. J. Zhang, Q. Qian, M. Cui, C. Chen, S. Liu and B. Han, Green Chem., 2017, 19, 4396–4401.
10. G. Chen, S. Li, F. Jiao and Q. Yuan, Catal. Today, 2007, 125, 111–119.
11. M. M. Doheim, S. A. Hanafy and G. A. El-Shobaky, Mater. Lett., 2002, 55, 304–311.
12. H. Knözinger, H. Bühl and K. Kochloefl, J. Catal., 1972, 24, 57–68.
13. B. C. Shi and B. H. Davis, J. Catal., 1995, 157, 359–367.
14. N. Takezawa, C. Hanamaki and H. Kobayashi, J. Catal., 1975, 38, 101–109.
15. X. Zhang, R. Wang, X. Yang and F. Zhang, Microporous Mesoporous Mater., 2008, 116, 210–215.
16. V. V. Bokade and G. D. Yadav, Appl. Clay Sci., 2011, 53, 263–271.
17. A. Ciftci, D. Varisli, K. Cem Tokay, N. Aslı Sezgi and T. Dogu, Chem. Eng. J., 2012, 207–208, 85–93.
18. J. Gurgul, M. Zimowska, D. Mucha, R. P. Socha and L. Matachowski, J. Mol. Catal. A: Chem., 2011, 351, 1–10.
19. T. Okuhara, T. Arai, T. Ichiki, K. Y. Lee and M. Misono, J. Mol. Catal., 1989, 55, 293–301.
20. Y. Saito and H. Niiyama, J. Catal., 1987, 106, 329–336.
21. A. T. Aguayo, A. G. Gayubo, A. Atutxa, M. Olazar and J. Bilbao, Ind. Eng. Chem. Res., 2002, 41, 4216–4224.
22. D. E. Bryant and W. L. Kranich, J. Catal., 1967, 8, 8–13.
23. R. Le Van Mao, P. Levesque, G. McLaughlin and L. H. Dao, Appl. Catal., 1987, 34, 163–179.
24. R. Le Van Mao, T. M. Nguyen and G. P. McLaughlin, Appl. Catal., 1989, 48, 265–277.
25. F. F. Madeira, N. S. Gnep, P. Magnoux, S. Maury and N. Cadran, Appl. Catal., A, 2009, 367, 39–46.
26. W. R. Moser, C.-C. Chiang and R. W. Thompson, J. Catal., 1989, 115, 532–541.
27. J. Ouyang, F. Kong, G. Su, Y. Hu and Q. Song, Catal. Lett., 2009, 132, 64–74.
28. C. B. Phillips and R. Datta, Ind. Eng. Chem. Res., 1997, 36, 4466–4475.
29. K. Ramesh, Y. L. E. Goh, C. G. Gwie, C. Jie, T. J. White and A. Borgna, J. Porous Mater., 2012, 19, 423–431.
30. K. Ramesh, L. M. Hui, Y.-F. Han and A. Borgna, Catal. Commun., 2009, 10, 567–571.
31. K. Ramesh, C. Jie, Y.-F. Han and A. Borgna, Ind. Eng. Chem. Res., 2010, 49, 4080–4090.
32. I. Takahara, M. Saito, M. Inaba and K. Murata, Catal. Lett., 2005, 105, 249–252.
33. N. Zhan, Y. Hu, H. Li, D. Yu, Y. Han and H. Huang, Catal. Commun., 2010, 11, 633–637.
34. D. Zhang, R. Wang and X. Yang, Catal. Lett., 2008, 124, 384–391.
35. T. K. Phung and G. Busca, Chem. Eng. J., 2015, 272, 92–101.
36. T. K. Phung, L. Proietti Hernández, A. Lagazzo and G. Busca, Appl. Catal., A, 2015, 493, 77–89.
37. M. Inaba, K. Murata, M. Saito and I. Takahara, React. Kinet. Catal. Lett., 2006, 88, 135–141.
38. K. A. Tarach, J. Tekla, W. Makowski, U. Filek, K. Mlekodaj, V. Girman, M. Choi and K. Gora-Marek, Catal. Sci. Technol., 2016, 6, 3568–3584.
39. M. Choi, H. S. Cho, R. Srivastava, C. Venkatesan, D.-H. Choi and R. Ryoo, Nat. Mater., 2006, 5, 718–723.
40. W. Fan, M. A. Snyder, S. Kumar, P.-S. Lee, W. C. Yoo, A. V. McCormick, R. Lee Penn, A. Stein and M. Tsapatsis, Nat. Mater., 2008, 7, 984–991.
41. B. T. Holland, L. Abrams and A. Stein, J. Am. Chem. Soc., 1999, 121, 4308–4309.
42. C. J. H. Jacobsen, C. Madsen, J. Houzvicka, I. Schmidt and A. Carlsson, J. Am. Chem. Soc., 2000, 122, 7116–7117.
43. Y. Tao, H. Kanoh and K. Kaneko, J. Am. Chem. Soc., 2003, 125, 6044–6045.
44. H. Wang and T. J. Pinnavaia, Angew. Chem., Int. Ed., 2006, 45, 7603–7606.
45. D. P. Serrano, J. Aguado, J. M. Escola, J. M. Rodríguez and Á. Peral, Chem. Mater., 2006, 18, 2462–2464.
46. M. Opanasenko, M. Shamzhy, F. Yu, W. Zhou, R. E. Morris and J. Cejka, Chem. Sci., 2016, 7, 3589–3601.
47. T. C. Keller, S. Isabettini, D. Verboekend, E. G. Rodrigues and J. Perez-Ramirez, Chem. Sci., 2014, 5, 677–684.
48. M. Choi, K. Na, J. Kim, Y. Sakamoto, O. Terasaki and R. Ryoo, Nature, 2009, 461, 246–249.
49. D. Xu, G. R. Swindlehurst, H. Wu, D. H. Olson, X. Zhang and M. Tsapatsis, Adv. Funct. Mater., 2014, 24, 201–208.
50. K. Rodponthukwaji, C. Wattanakit, T. Yutthalekha, S. Assavapanumat, C. Warakulwit, W. Wannapakdee and J. Limtrakul, RSC Adv., 2017, 7, 35581–35589.
51. W. Wannapakdee, C. Wattanakit, V. Paluka, T. Yutthalekha and J. Limtrakul, RSC Adv., 2016, 6, 2875–2881.
52. P. Wuamprakhon, C. Wattanakit, C. Warakulwit, T. Yutthalekha, W. Wannapakdee, S. Ittisanronnachai and J. Limtrakul, Microporous Mesoporous Mater., 2016, 219, 1–9.
53. T. Yutthalekha, D. Suttipat, S. Salakhum, A. Thivasasith, S. Nokbin, J. Limtrakul and C. Wattanakit, Chem. Commun., 2017, 53, 12185–12188.
54. X. Zhang, D. Liu, D. Xu, S. Asahina, K. A. Cychosz, K. V. Agrawal, Y. Al Wahedi, A. Bhan, S. Al Hashimi, O. Terasaki, M. Thommes and M. Tsapatsis, Science, 2012, 336, 1684–1687.
55. A. J. J. Koekkoek, C. H. L. Tempelman, V. Degirmenci, M. Guo, Z. Feng, C. Li and E. J. M. Hensen, Catal. Today, 2011, 168, 96–111.
56. E. M. Flanigen, J. M. Bennett, R. W. Grose, J. P. Cohen, R. L. Patton, R. M. Kirchner and J. V. Smith, Nature, 1978, 271, 512–516.
57. B. Boekfa, P. Pantu, M. Probst and J. Limtrakul, J. Phys. Chem. C, 2010, 114, 15061–15067.
58. T. Maihom, S. Wannakao, B. Boekfa and J. Limtrakul, Chem. Phys. Lett., 2013, 556, 217–224.
59. W. Panjan, J. Sirijaraensre, C. Warakulwit, P. Pantu and J. Limtrakul, Phys. Chem. Chem. Phys., 2012, 14, 16588–16594.
60. C. E. Hernandez-Tamargo, A. Roldan and N. H. de Leeuw, J. Solid State Chem., 2016, 237, 192–203.
61. D. Zhai, L. Zhao, Y. Liu, J. Xu, B. Shen and J. Gao, Chem. Mater., 2015, 27, 67–74.
62. B. Boekfa, S. Choomwattana, P. Khongpracha and J. Limtrakul, Langmuir, 2009, 25, 12990–12999.
63. T. Maihom, B. Boekfa, J. Sirijaraensre, T. Nanok, M. Probst and J. Limtrakul, J. Phys. Chem. C, 2009, 113, 6654–6662.
64. S. K. Jesudoss, J. J. Vijaya, K. Kaviyarasu, L. J. Kennedy, R. Jothi Ramalingam and H. A. Al-Lohedan, RSC Adv., 2018, 8, 481–490.
65. M. Liu, J. Li, W. Jia, M. Qin, Y. Wang, K. Tong, H. Chen and Z. Zhu, RSC Adv., 2015, 5, 9237–9240.
66. O. Ben Mya, M. Bita, I. Louafi and A. Djouadi, MethodsX, 2018, 5, 277–282.
67. N. Katada, H. Igi and J.-H. Kim, J. Phys. Chem. B, 1997, 101, 5969–5977.
68. S. Kouva, J. Kanervo, F. Schüβler, R. Olindo, J. Lercher and A. Krause, Chem. Eng. Sci., 2013, 89, 40–48.
69. D. Ma, Y. Lu, L. Su, Z. Xu, Z. Tian, Y. Xu, L. Lin and X. Bao, J. Phys. Chem. B, 2002, 106, 8524–8530.
70. S. Li, A. Zheng, Y. Su, H. Zhang, L. Chen, J. Yang, C. Ye and F. Deng, J. Am. Chem. Soc., 2007, 129, 11161–11171.
71. S. Li, S.-J. Huang, W. Shen, H. Zhang, H. Fang, A. Zheng, S.-B. Liu and F. Deng, J. Phys. Chem. C, 2008, 112, 14486–14494.
72. C. Liu, G. Li, E. J. M. Hensen and E. A. Pidko, ACS Catal., 2015, 5, 7024–7033.
73. R. Ravishankar, C. E. A. Kirschhock, P.-P. Knops-Gerrits, E. J. P. Feijen, P. J. Grobet, P. Vanoppen, F. C. De Schryver, G. Miehe, H. Fuess, B. J. Schoeman, P. A. Jacobs and J. A. Martens, J. Phys. Chem. B, 1999, 103, 4960–4964.
74. G. Engelhardt, U. Lohse, A. Samoson, M. Mägi, M. Tarmak and E. Lippmaa, Zeolites, 1982, 2, 59–62.
75. D. Varisli, T. Dogu and G. Dogu, Chem. Eng. Sci., 2007, 62, 5349–5352.
76. K. K. Ramasamy and Y. Wang, Catal. Today, 2014, 237, 89–99.
77. Z. S. B. Sousa, C. O. Veloso, C. A. Henriques and V. Teixeira da Silva, J. Mol. Catal. A: Chem., 2016, 422, 266–274.
78. M. A. Christiansen, G. Mpourmpakis and D. G. Vlachos, ACS Catal., 2013, 3, 1965–1975.
79. M. Zhang and Y. Yu, Ind. Eng. Chem. Res., 2013, 52, 9505–9514.
80. W. Choopun and S. Jitkarnka, J. Cleaner Prod., 2016, 135, 368–378.
81. C.-Y. Wu and H.-S. Wu, ACS Omega, 2017, 2, 4287–4296.
82. M. Hartmann, A. G. Machoke and W. Schwieger, Chem. Soc. Rev., 2016, 45, 3313–3330.
83. H. Zhang, Z. Hu, L. Huang, H. Zhang, K. Song, L. Wang, Z. Shi, J. Ma, Y. Zhuang, W. Shen, Y. Zhang, H. Xu and Y. Tang, ACS Catal., 2015, 5, 2548–2558.
84. R. Jérôme, R. Pascal and C. Céline, ChemCatChem, 2017, 9, 2176–2185.
85. H. Chiang and A. Bhan, J. Catal., 2010, 271, 251–261.
86. P. M. Mortensen, J. D. Grunwaldt, P. A. Jensen, K. G. Knudsen and A. D. Jensen, Appl. Catal., A, 2011, 407, 1–19.
87. F. Sannino, M. Pansini, A. Marocco, B. Bonelli, E. Garrone and S. Esposito, Phys. Chem. Chem. Phys., 2015, 17, 28950–28957.
88. K. Alexopoulos, M. John, K. Van der Borght, V. Galvita, M.-F. Reyniers and G. B. Marin, J. Catal., 2016, 339, 173–185.
89. S. A. Kadam and M. V. Shamzhy, Catal. Today, 2018, 304, 51–57.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c8se00392k

‡ These authors contributed equally.

---