The Influence of Mesopore Architecture in Hierarchical H‑ZSM‑5 on n‑Butanol Dehydration

Abstract

Zeolites are among the most widely employed catalysts in the (petro-)­chemical industry. However, due to their elaborate microporous network, they are prone to diffusion limitations and deactivation. Several modification methods have been proposed to overcome these limitations, each exhibiting their benefits. In this work, two of the most promising strategies were combined, i.e., limiting the length of one of the crystal axes during synthesis to achieve a platelike morphology and introducing mesoporosity, creating a hierarchical platelike H-ZSM-5. The platelike morphology was obtained by adding urea as a growth modifier to the synthesis mixture, and mesopores were introduced in the platelike H-ZSM-5 through etching with a NaOH/TPAOH mixture. As a benchmark, the same etching procedure was applied to a commercial ZSM-5 counterpart. These materials were tested in the n-butanol dehydration, where the platelike morphology exhibited an improved catalytic performance, significantly increasing the activity per acid site and stability, and slightly increasing the selectivity toward the butenes. The generation of mesopores in commercial ZSM-5 also increased the activity per acid site but reduced the catalyst’s stability, likely due to an increased amount of Lewis acid sites upon etching. When applying the same modification method to the platelike H-ZSM-5, much larger mesopores and some macropores were observed. These further increased the stability of the catalyst but barely affected the activity per acid site, presumably due to the already optimized catalytic performance of the platelike H-ZSM-5.

1. Introduction

Since fossil fuels are finite and harm the environment, it is important to find alternative sustainable and renewable resources. A widely investigated and already implemented sustainable production route is the fermentation of biomass to bioalcohols. , These bioalcohols, including n-butanol, are already used as sustainable fuels, and due to their increasing importance, biobutanol production rises and a surplus is created. This makes biobutanol an interesting green resource for several platform chemicals such as butenes, which can be formed through catalytic dehydration of n-butanol using H-ZSM-5 as a catalyst.

The dehydration of n-butanol over H-ZSM-5 can proceed through different direct and indirect reaction paths, each consisting of multiple reaction mechanisms. − It is assumed that 1-butene can be formed via the direct dehydration of n-butanol or an indirect reaction path with dibutyl ether as an intermediate. The 2-butene isomers can also be formed through the direct and indirect dehydration of n-butanol, but also through double bond isomerization reactions. Lastly, isobutene is formed through skeletal isomerization of the linear butenes, but only at temperatures above 573 K. − The n-butanol dehydration has been studied extensively within our group, and more information about the reaction mechanism can be found in earlier work. −

Although H-ZSM-5 exhibits high activity and selectivity for the dehydration of butanol to butenes, it also presents some drawbacks, such as diffusion limitations and deactivation due to coke formation. , To overcome these drawbacks, several possible solutions are proposed, such as zeolites with a special morphology and hierarchical zeolites. One of the most promising strategies is controlling the zeolite morphology by limiting the growth along one of the zeolite’s channels during synthesis. By doing so, it is possible to bring one of its planes to the external surface to increase the number of pore mouths per unit crystal. For ZSM-5, which has straight channels along the b-axis and sinusoidal channels along the a-axis, limiting the growth along the b-axis has improved the catalytic performance in several reactions such as benzene alkylation, , catalytic cracking, , and methanol-to-propylene. − Reducing the b-axis reduces the residence time inside these straight channels and improves the transport efficiency to and away from the active sites. This results in a more effective utilization of the acid sites, increasing the conversion. − , In addition, products are more easily removed, reducing the probability of interaction with multiple acid sites within the channels, i.e., secondary reactions toward more condensed products or coke precursors and coke species themselves. This shifts selectivities toward lighter molecules and increases the catalyst’s stability by decreasing the coke formation rate and reducing coke deposition. − ,, By increasing the number of pore mouths on the surface, the accessibility of the acid sites is improved, further enhancing the conversion. This also results in a superior coke tolerance as the catalyst can contain a higher coke content before affecting the activity, thus further slowing down deactivation. −

In 2009, Ryoo et al. pioneered by synthesizing nanosheet ZSM-5 with a b-axis length limited to 2 nm, i.e., the dimension of the b-axis in one single unit cell. These zeolites had a large number of acid sites on the external surface, resulting in a highly active catalyst, especially for reactions involving large organic molecules such as the methanol-to-gasoline reaction. Since then, several other synthesis methods for nanosheet ZSM-5 have been proposed, but they often require complex, expensive, and environmentally unfriendly structure-directing agents and/or a long synthesis procedure, rendering them undesirable for upscaling as it would encounter high energy and product costs. ,, The synthesis process can be significantly simplified by limiting the length of the b-axis to 10–100 nm instead of the dimensions of the unit cell. Recently, a synthesis method using NH₄F has been developed to synthesize platelike ZSM-5 with a b-axis of ca. 100 nm. ,, This catalyst exhibited excellent results in the methanol-to-hydrocarbons reaction, increasing its activity and extending its lifetime. However, due to the health, safety and environmental issues concerning fluoride streams, scale-up could be limited. A greener alternative is adding a growth modifier to the synthesis gel. These growth modifiers can interact with a specific facet of the zeolite crystal and/or interact with amorphous precursors, regulating the growth kinetics to change the crystal size and morphology. Among them, urea is often used as it is green, cheap and effective. , Urea spontaneously adsorbs on zeolite surfaces, , preferentially on the ac-plane, therefore hindering the growth along the b-axis.

Another promising strategy involves hierarchical zeolites, which differ from conventional zeolites by incorporating an additional porous network, typically meso- (2–50 nm) or macroporous (>50 nm), interconnecting the micropores (<2 nm) of the zeolite. These meso- (or macro-)­pores can occur inside the microporous crystals, i.e., intracrystalline meso- (or macro-)­porosity, or in between crystal boundaries, i.e., intercrystalline meso- (or macro-)­porosity, and enhance the accessibility of the acid sites, increase the specific surface area, and reduce the diffusion path length. ,,

Hierarchization is performed during or after synthesis, referred to as bottom-up or top-down approaches, respectively. The former generally makes use of additional templates to form the secondary porosity, i.e., the meso- or macropores, and removal of these templates upon calcination or dissolution after crystallization is required. − Top-down approaches consist of dealumination and desilication or a combination of both. During dealumination Al–O–Si bonds are broken through acid, steam or heat treatments, releasing Al and creating mesopores. However, the formed pores tend to have a wide pore size distribution and are not interconnected with one another, limiting their benefits. Furthermore, due to the selective extraction of Al, the acidity of the catalyst decreases. , Desilication, also known as base-leaching, with NaOH, is the most well-known hierarchization method for ZSM-5 as it is a simple, effective, and economical strategy. Here, Si–O–Si bonds are hydrolyzed, releasing preferentially Si from the zeolite framework to create mesopores. , Nevertheless, to obtain a successful hierarchization, Si/Al ratios of the parent material have to be limited within the 25–50 range. When using lower Si/Al ratios, hydrolysis of the Si-atoms will be prevented by the surrounding Al-atoms in the framework, creating only a limited amount of mesopores. For higher Si/Al ratios, on the other hand, desilication will be uncontrollable, forming large pores or even fully dissolving the zeolite framework. To enable desilication on a wider range of zeolites, several variations have been proposed.

Qin et al. used an HF solution to retain the initial Si/Al ratio and microporous structure since HF is known to react with both Si and Al. However, the concentration of the solution must be sufficiently high because a diluted solution primarily extracts Al. Adding NH₄F to the HF etching solution allows more diluted solutions to create hierarchical ZSM-5 with large meso- and macropores and with the intrinsic micropores preserved. , However, due to the dangerous nature of HF, an etching method with solely NH₄F was created. − Here, HF is formed in situ, reducing the HF-concentration and making it a safer method. In addition, the method was more efficient and controllable than the one using HF or an HF/NH₄F mixture. Earlier research within our group demonstrated the positive effects of the mesopores generated by NH₄F-etching on the n-butanol dehydration. The etching targeted intergrowths in aggregates, cutting these aggregated crystals from the edge toward the center and creating intercrystalline mesoporosity. When using parent zeolites with a less aggregated structure, defective sites within the zeolites are attacked, etching rectangular nanodomains, thus creating intracrystalline mesoporosity. Catalytic testing in the n-butanol dehydration revealed an increased activity and selectivity for the etched materials, overcoming an activity-stability trade-off where a material typically becomes less stable when the activity increases.

Although this method presents interesting perspectives, it still relies on fluoride components, with their inherent safety issues for scale-up. However, the standard NaOH desilication mixture can also be modified to overcome the Si/Al ratio limits. Pérez-Ramírez et al. created a modification method where tetrapropylammonium hydroxide (TPAOH) is used as a pore-growth moderator in the desilication with NaOH. During modification, both Na⁺ and TPA⁺ competitively adsorb on the zeolite surface to balance the negative charges. The large size and branched structure of the TPA⁺ ion protects the neighboring Si, preventing its extraction. Since Na⁺ does not have this protective effect, desilication mainly occurs near Na⁺ cations. This protective effect of TPA⁺ makes it a highly controllable desilication method, as changing the concentration of the two cations in the mixture can change the desilication severity and makes it possible to create a hierarchical zeolite with small, uniformly distributed mesopores throughout the whole crystal. , These uniformly distributed mesopores have shown great results in several reactions, including the benzene alkylation, the catalytic upgrading of bio-oil, and the methanol-to-hydrocarbon process. The secondary porosity improves the molecular transport to and from the active sites, thus enhancing the catalyst’s activity. Furthermore, there is a higher coke tolerance, and the removal of coke precursors from the zeolite channels is enhanced, decreasing the coking rate.

Despite the extensive research on ZSM-5, the comparative analysis of the different structural modifications of ZSM-5 under identical reaction conditions remains scarce. Additionally, research on the application of the combination of both modification methods, i.e., hierarchical platelike H-ZSM-5, which combines the advantageous properties of both the short b-axis and the secondary porosity, is limited. In our present work, at first the effect of the H-ZSM-5 morphology on its performance in the n-butanol dehydration is examined. Platelike H-ZSM-5 with Si/Al of 25 is synthesized through hydrothermal synthesis, adding urea to inhibit the growth along the b-axis. Subsequently, an assessment is done of the effect of the catalyst’s morphology on the hierarchization severity of the desilication with a NaOH and TPAOH mixture. This modification method is applied to commercial H-ZSM-5 with a Si/Al of 25 and the self-synthesized platelike H-ZSM-5, also with Si/Al of 25. All zeolites are fully characterized by XRD, SEM, TEM, ICP-OES, N₂ sorption, FTIR, NH₃-TPD, and ²⁷Al NMR, and tested in the n-butanol dehydration reaction on a high-throughput kinetics screening setup at temperatures between 483 and 513 K to enable a comprehensive understanding of the effects of the morphology and mesopores on the catalyst’s activity, selectivity and stability.

2. Materials and Methods

2.1. Zeolite Materials

Platelike H-ZSM-5 with a Si/Al ratio of 25 (referred to as PL-ZSM-5) is synthesized using a modified method from Qureshi et al. and Li et al. First, tetraethyl orthosilicate (TEOS, >97.0%, TCI) is added dropwise to H₂O under continuous stirring at room temperature. This is followed by the dropwise addition of the structure directing agent tetrapropylammonium hydroxide (TPAOH, 1.0 M in water, Sigma-Aldrich). When a clear solution is obtained, the Al source (Al­(NO₃)₃·9H₂O, ≥98.5%, ChemLab) is added, followed by NaOH (≥99%, ChemLab) and isopropyl alcohol (99.8%). Lastly, urea (≥99.5%, Merck) is added to limit the growth along the b-axis, forming a platelike morphology. The synthesis mixture with a molar ratio of 472 H₂O: 25 Si: 7.55 TPAOH: 1 Al: 1.18 NaOH: 0.75 IPA: 14.15 urea is hydrolyzed at room temperature for 6 h under continuous stirring at 500 rpm. Afterward, the mixture is transferred into a Teflon-lined stainless-steel autoclave for hydrothermal reaction at 453 K for 48 h. After crystallization, the solid product is recovered by centrifugation and washed with deionized water until the pH of the washing water becomes neutral. Finally, the synthesized zeolite is dried at 373 K for 24 h followed by calcination in air at 823 K for 5 h with a temperature ramp of 1 K min^–1 to remove the template.

To bring the synthesized zeolite to its acid form, it is ion-exchanged in a 1 M solution of NH₄NO₃ (≥98%, Sigma-Aldrich) in water under continuous stirring at 323 K for 2 h. Next, the zeolite is separated from the liquid by centrifugation, followed by a washing step with deionized water at room temperature. The ion-exchange is repeated three times to obtain the NH₄ ⁺-form of the zeolite. The H⁺-form is then obtained upon calcination at 823 K for 4 h with a temperature ramp of 1 K min^–1.

Both commercial (Zeolyst, CBV 5524 G, Si/Al ratio = 25, referred to as c-ZSM-5) and platelike H-ZSM-5 zeolites are modified through desilication in an aqueous base solution of 0.2 M NaOH and 0.25 M TPAOH, based on a method described by Chaihad et al. For this, 2 g parent zeolite is added to a flask of DURAN borosilicate glassware together with 10 mL 0.2 M NaOH and 10 mL 0.25 M TPAOH and placed in a preheated oil bath at 353 K, where it is continuously stirred for 1 h under reflux conditions. Subsequently, the desilication is stopped by immediately quenching the mixture in an ice–water bath, after which the catalyst is separated from the liquid by centrifugation. This is followed by washing with deionized water at room temperature until the washing water is neutral, and finally dried at 373 K overnight. After desilication, a mass loss of 13% was observed for c-ZSM-5 and 14% for PL-ZSM-5. This includes both framework extraction and minor losses during washing and handling. The hierarchical zeolites are transformed into their NH₄ ⁺-form by three ion-exchanges in a 1 M solution of NH₄NO₃ in water under continuous stirring for 2 h at 323 K. This is followed by a drying step at 373 K for 24 h and calcination at 823 K for 4 h with a temperature ramp of 1 K min^–1 to obtain its H⁺-form.

2.2. Zeolite Characterization

X-ray diffraction (XRD) assesses the zeolite structure and whether structural changes occur after desilication. For this, diffraction patterns with 2θ between 5° and 60° are collected (Bruker D8 Advance diffractometer with Cu Kα radiation). The zeolites’ surface areas and pore volumes are obtained through N₂ sorption at 77 K (Micromeritics Tristar). Before the measurements, the catalysts are pretreated under continuous N₂ flow at 573 K for 4 h to remove moisture and adsorbed species. The BET method is used to determine the surface area (S _BET), the t-plot method for the external surface area (S _ext), micropore surface area (S _micro) and micropore volume (V _micro), and the total pore volume (V _tot) is obtained through single point adsorption. To increase the accuracy of the BET surface area, Rouquerol’s criteria are applied. Pore size distributions are calculated using the Barrett–Joyner–Halenda (BJH) method on the adsorption branch.

More insight into morphology, crystal size and crystal size distribution is obtained through transmission electron microscopy (TEM, JEOL JEM-2200FS), using Cu and Ni supports, and scanning electron microscopy (SEM, FEI Quanta 200 F). Furthermore, inductively coupled plasma optical emission spectroscopy (ICP-OES) is used to verify the Si/Al ratio of the zeolites. Before analysis, the sample is prepared by destruction in an HF solution, which is analyzed using a Thermo Scientific IcaP 7400 ICP-OES.

Solid state ²⁷Al MAS NMR is performed on a Bruker AVANCE 500 NB spectrometer, using a rotational speed of 12 kHz. The experiments are performed at 130.30 MHz with a 4 mm probe head, and Al­(NO₃)₃ as a reference for the chemical shift. A total of 2048 scans are collected, using a π/6 pulse with a radiofrequency field strength of 107 kHz and a recycle delay of 1 s. Prior to the measurements, the samples are kept overnight at a relative humidity of 75% using a saturated NaCl solution. Silanol groups and acid sites are assessed through Fourier transform infrared (FTIR) using a Nicolet Magna 550-FT-IR in transmission mode. Before analysis, the samples are pressed in self-supported discs with a diameter of 16 mm and in situ degassed for 4 h at room temperature under vacuum. Quantification of the Brønsted and Lewis acid sites is done by in situ monitoring of pyridine adsorption and desorption at 473 K. The FTIR results are normalized to a pellet density of 10 mg/cm² and the amount of Brønsted acid sites (BAS) is calculated from the area of the band at 1545 cm^–1, using an extinction coefficient ε­(B)₁₅₄₅ = 1.67 cm/μmol, and for the Lewis acid sites (LAS) from the band at 1454 cm^–1 with the extinction coefficient ε­(L)₁₄₅₄ = 2.22 cm/μmol. The strength of the acid sites was also determined using NH₃-temperature-programmed desorption (NH₃-TPD, Micromeritics Autochem). The sample is saturated with NH₃ by flowing a 4 mol % NH₃/He-flow at 423 K and unbound and/or physisorbed NH₃ is removed by flushing He over the sample. The NH₃ is desorbed by heating the sample to 923 K at a rate of 10 K min^–1, the amount of NH₃ that desorbs is tracked with a thermal conductivity detector (TCD) and mass spectrometer (MS). Afterward, the NH₃-TPD profiles are deconvoluted into three peaks corresponding to weak (423–523 K), medium (523–623 K), and strong (>623 K) acid sites. More details about the deconvolution can be found in the Supporting Information.

2.3. Zeolite Testing

The catalyst samples are prepared by pelletizing and sequential sieving until particle sizes within a 75–100 μm range are obtained, followed by drying for 24 h to remove adsorbed water. The dried catalyst is weighted (mass range of 0.010–0.050 g) and diluted with inert α-Al₂O₃ until 10 mass% to avoid temperature deviations across the catalyst bed. The catalytic experiments are performed in a high-throughput kinetics screening setup comprising four reactor blocks, each containing four tubular reactors with a length of 80 cm and an inner diameter of 2.1 mm. The catalyst is packed in the center of the reactor between layers of inert α-Al₂O₃. Liquid n-butanol (>99.5%, Merck, referred to as BuOH) is used as feedstock and nitrogen (Air Liquide) as an inert carrier gas to regulate the partial pressure of the feed to 29 kPa at reactor inlet, similar to previous research. ,,− Before each experiment, the catalyst bed is heated under an N₂ flow for at least 4 h to ensure stable reaction temperatures. To avoid the condensation of heavier products in the effluent, all lining downstream the reactor section is heated to 443 K. Online analysis of the effluent is performed with a GC-FID (Thermo Fisher Scientific Trace 1310 with an HP-PONA column) with ethane as internal standard to calculate the flows and product distributions, and to verify the mass balances. Although the FID-detector cannot detect water, the amount of water that is formed during the reaction is calculated using the reaction stoichiometry since no oxygenated hydrocarbons apart from butanol and dibutyl ether are detected, indicating all oxygen is removed in the form of water. A time on stream (TOS) of zero is defined as the start of the butanol feed. All measurements in this work are performed between one and 10 h TOS to ensure a stable operating regime. Lastly, deactivation tests are performed for 60 h time on stream. All experiments show a closed carbon balance within 5%. Error bars depicted in the data correspond to the 95% confidence intervals, which are derived from repeating experiments under identical conditions.

Since the different zeolites studied do not have the same amount of active sites per gram of catalyst, the activity is normalized based on the amount of active sites. To this end, the activity is not expressed per space time of catalyst weight, but per site time (SiT), which is the number of active sites present during reaction divided by the molar flow of the butanol feed, therefore indicative for the average contact time between reactant and active site

$$\mathrm{SiT}=\frac{W·C_{\mathrm{a}}}{F_{\mathrm{BuOH}}^0}$$ 1

with W the mass of the catalyst [kg], C _a the concentration of the active sites [mol kg^–1], here defined as the number of Brønsted acid sites determined with pyridine FTIR, and F _BuOH the molar flow of n-butanol fed to the reactor [mol s^–1]. As such, site time is expressed in seconds. The activity of the catalysts is based on the conversion of n-butanol (X_BuOH), which is defined as

$$X_{\mathrm{BuOH}}=\frac{F_{\mathrm{BuOH}}^0-F_{\mathrm{BuOH}}}{F_{\mathrm{BuOH}}^0}$$ 2

with F _BuOH the molar flow of n-butanol at the reactor outlet [mol s^–1], determined upon GC analysis. The activity can also be compared in terms of turnover frequency (TOF), which is the amount of converted reagent per active site of the catalyst, per unit of time

$$\mathrm{TOF}=\frac{F_{\mathrm{BuOH}}^0-F_{\mathrm{BuOH}}}{W·C_{\mathrm{a}}}$$ 3

The carbon selectivity toward products i (S_i ) is expressed as

$$S_i=\frac{b_i{F}_i}{c_{\mathrm{BuOH}}{F}_{\mathrm{BuOH}}^0}$$ 4

where b _i is the number of carbon atoms present in component i, F _i the molar flow of component i at the reactor outlet [mol s^–1] and c _BuOH the number of carbon atoms in the butanol feed (= 4).

3. Results and Discussion

3.1. Characterization

The XRD patterns of the commercial and synthesized catalysts and their modified versions are shown in Figure and exhibit the characteristic peaks of the MFI topology between 7.5 and 9.5° and between 22.5 and 25° for all samples, confirming the successful synthesis of the MFI structure in platelike ZSM-5 (PL-ZSM-5). Furthermore, it illustrates that no structural changes occur upon modification of the commercial ZSM-5 (m-ZSM-5) or platelike ZSM-5 (m-PL-ZSM-5), as no signals associated with impurities are detected. The relative crystallinity of the hierarchical and platelike samples with respect to the commercial one is determined by comparing the sum of the areas of the XRD patterns between 2θ = 7.5–9.5° and 22.5–25° to those areas in c-ZSM-5. ,,− Both m-ZSM-5 and PL-ZSM-5 have a relative crystallinity of 85%, and m-PL-ZSM-5 has one of 91%, indicating only a minor impact on the crystallinity.

1.

XRD patterns of c-ZSM-5 (a), m-ZSM-5 (b), PL-ZSM-5 (c), and m-PL-ZSM-5 (d).

Catalyst morphology, particle size and aggregation of crystals are assessed through SEM. The commercially available H-ZSM-5 (c-ZSM-5) has small, irregularly shaped crystals that aggregate into larger structures as shown in Figure a. Furthermore, they exhibit a wide particle size distribution, extending from about 60 to 600 nm. The self-synthesized PL-ZSM-5, see Figure e, has coffin-shaped crystals with reduced b-axes of approximately 70 nm and c-axes between 700 and 1400 nm. Some smaller crystals have a nicely finished hexagonal coffin shape on one side but a rough edge on the other, indicating that some crystals are broken in half with resulting c-axes between 400 and 700 nm. The SEM images of m-ZSM-5 and m-PL-ZSM-5 are similar to those of their parent materials (Figure S1), as expected that the addition of secondary porosity does not alter the shape and size of the crystals.

2.

SEM of c-ZSM-5 (a), TEM of c-ZSM-5 (b), TEM of m-ZSM-5 (c, d), SEM of PL-ZSM-5 (e), TEM of PL-ZSM-5 (f), and TEM of m-PL-ZSM-5 (g, h). The crack-like features observed on the crystal surfaces in the SEM images are attributed to the gold coating applied to discharge the sample and enhance the resolution.

To assess the mesoporosity, N₂ sorption is performed, and the isotherms are presented in Figure (and Figure S2). The c-ZSM-5 isotherm is of the type I­(a) according to the IUPAC classification, which is characteristic of microporous materials. Furthermore, the isotherm has a small hysteresis of type H4, which can be ascribed to the presence of a small amount of intercrystalline mesopores and/or crystal aggregates. , Due to the pelletizing of the catalyst particles, some minor intercrystalline and/or interparticle macroporosity is noticeable. For m-ZSM-5, the isotherm can be classified as a type IV­(a) with a hysteresis with type H2 and H4 characteristics. The pronounced hysteresis at lower relative pressures, resembling type H2, indicates the generation of intracrystalline ink bottle-shaped mesopores after modification, whereas the H4 type hysteresis at higher relative pressures resembles slit-shaped meso- and macropores. ,,− Similarly to c-ZSM-5, PL-ZSM-5 shows an I­(a) type isotherm, with a smaller type H4 hysteresis, which is mainly present at higher partial pressures, indicating that the meso- and/or macropores formed between the crystals are larger. Desilication of PL-ZSM-5 also results in a type IV­(a) isotherm with a type H2 and H4 hysteresis, similar to m-ZSM-5. Nevertheless, the hysteresis of m-PL-ZSM-5 is more open toward higher relative pressures (P/P⁰ > 0.8), indicating larger mesopores than m-ZSM-5.

3.

N₂ sorption isotherms for c-ZSM-5 (a, red), m-ZSM-5 (a, dashed red), PL-ZSM-5 (b, blue), and m-PL-ZSM-5 (b, dashed blue).

Specific surface areas and pore volumes are shown in Table . Both parent zeolites exhibit similar micropore surface areas (S _micro, 395 cm³ g^–1 for c-ZSM-5 and 390 cm³ g^–1 for PL-ZSM-5) and micropore volumes (V _micro, both 0.18 cm³ g^–1). However, the external surface area (S _ext) and mesopore volume (V _meso) are higher for c-ZSM-5, namely 46 m² g^–1 and 0.07 cm³ g^–1, compared to 25 m² g^–1 and 0.05 cm³ g^–1 for PL-ZSM-5. The t-plot method, which is used to determine the external surface area, does not differentiate between the external surface area and mesopore surface area. This higher external surface area is thus likely related to the presence of a small amount of intercrystalline mesopores in c-ZSM-5. , The desilication of c-ZSM-5 results in an increased BET surface area and total pore volume due to the additional mesoporous surface area and volume, without compromising its micropore surface area and volume. It can, therefore, be assumed that there is no loss of microporosity after the modification of commercial H-ZSM-5 with a mixture of NaOH and TPAOH, making it a mild etching method compared to other desilication methods or dealumination. Similarly, the hierarchization of PL-ZSM-5 increases the mesopore surface area and volume. However, a decrease in BET surface area (8%), micropore surface area (14%) and volume (13%) is observed, indicating a small loss of microporosity.

1. Surface Areas and Pore Volumes Determined with N₂ Sorption.

catalyst	SBET [m2 g–1]	Sext [m2 g–1]	Smicro [m2 g–1]	Vtot [cm3 g–1]	Vmicro [cm3 g–1]	Vmeso [cm3 g–1]
c-ZSM-5	441	46	395	0.25	0.18	0.07
m-ZSM-5	468	75	393	0.30	0.17	0.13
PL-ZSM-5	415	25	390	0.23	0.18	0.05
m-PL-ZSM-5	380	44	336	0.25	0.15	0.10

The presence of larger pores in m-PL-ZSM-5 is corroborated by the BJH pore size distributions, as shown in Figure S3. The mesopores present in m-ZSM-5 vary between 5 and 25 nm with a modal diameter of 12 nm, whereas the secondary pore sizes in m-PL-ZSM-5 have a much broader distribution and vary from 5 nm to over 100 nm with a modal pore diameter of approximately 60 nm. The BJH method loses its accuracy for macropores and is only depicted until 100 nm.

Secondary porosity can also be visualized through TEM analysis for both hierarchical commercial H-ZSM-5 (m-ZSM-5) and hierarchical platelike H-ZSM-5 (m-PL-ZSM-5), as shown in Figure c,dand Figure g,h respectively. Measurements of the pores in all TEM images revealed small, intracrystalline mesopores between approximately 5 and 28 nm for m-ZSM-5, similar to the BJH pore size distribution (Figure S3). Whereas the secondary porosity of m-PL-ZSM-5 varies from mesopores to macropores, with the latter appearing to be formed by combining several smaller mesopores. Here, pore sizes between 6 and 350 nm are measured, showing the presence of very large cavities, which were not detected with the BJH pore size distribution as this method is optimized to measure mesopores.

The increased effect of desilication can not only be ascribed to the platelike morphology but may also be due to differences in the number of structural defects and the location of the Al sites in the synthesized PL-ZSM-5. The presence of extra-framework aluminum (EFAl) on the outer surface of the platelike crystals may prevent the leaching of Si-atoms at its surface. , Normalized FTIR spectra in the OH-stretching region in Figure show the presence of the silanols in the different samples. The peak at 3745 cm^–1 corresponds to isolated silanols located on the external surface of the zeolite , and increases by approximately 15% and 50% upon desilication of the commercial and platelike zeolite respectively, due to an increased external surface. Internal silanols typically show three different peaks at 3735, 3725, and 3700 cm^–1, corresponding to internal silanols in different environments. It is also possible that some silanol groups interact, forming silanol nests that show a broad band centered around 3500 cm^–1. , Upon desilication of both c-ZSM-5 and PL-ZSM-5, a decrease in the total amount of internal silanols (70% and 30%, respectively) and silanol nests (67% and 54%, respectively) is observed. This indicates that silicon extraction preferably occurs at these defective Si-sites. However, even though the total amount of internal silanols decreases, the peak around 3700 cm^–1 decreases the most, followed by the peak around 3725 cm^–1 and the peak around 3735 cm^–1 even slightly increases. This suggests that there is a shift in the type of internal silanols upon hierarchization, shifting from more constrained to more free. The band at 3665 cm^–1 can be attributed to the presence of extra-framework aluminum (EFAl) creating Lewis acidity, and the band at 3610 cm^–1 corresponds to bridging Si–OH-Al groups, providing Brønsted acidity. ,, These imply that the commercial material has a larger amount of Brønsted acid sites than the platelike ZSM-5 and that desilication affects the number of those Brønsted acid sites, while increasing the amount of EFAl. Deconvolution of the peaks shows that m-ZSM-5 has ca. 25% less BAS and 40% more EFAl than c-ZSM-5. For the platelike zeolites on the other hand, the number of BAS and EFAl increase with approximately 12% and 14%, respectively.

4.

Normalized FTIR in the OH-stretching region for c-ZSM-5 (red), m-ZSM-5 (dashed red), PL-ZSM-5 (blue), and m-PL-ZSM-5 (dashed blue) activated at 723 K. The peak at 3745 cm^–1 corresponds to the external silanols, the peaks at 3735, 3725 and 3700 cm^–1 correspond to the internal silanols, the peak at 3665 cm^–1 to the extra-framework aluminum, the peak at 3610 cm^–1 to the Brønsted acid sites, and the broad peak with a maximum around 3500 cm^–1 is correlated to the internal silanol nests.

Quantification of these Brønsted and Lewis acid sites through pyridine FTIR (Py-FTIR) is shown in Table together with the Si/Al ratio and Al-concentration obtained through ICP-OES. The Si/Al ratio of the synthesized PL-ZSM-5 of 29.2 is close to the desired ratio of 25.0, and after desilication, the Si/Al ratio and concentration of Al remain almost completely preserved, whereas, during conventional desilication processes, the Si/Al ratios tends to decrease more substantially since NaOH is more selective toward Si compared to TPAOH. , However, the number of acid sites measured with Py-FTIR is considerably lower for PL-ZSM-5, about half the amount as for the c/m-ZSM-5. Especially the Brønsted acid sites seem to be inaccessible for the pyridine, as the platelike exhibits less than half the amount of BAS compared to the commercial one with a similar Si/Al ratio. Etching with NaOH and TPAOH results in an increase in Lewis acidity for both parent catalysts, as already indicated by the FTIR results in the OH-stretching region, whereas the number of Brønsted acid sites decreases for m-ZSM-5 and remains unaltered for m-PL-ZSM-5. During alkaline treatment, distorted tetrahedral Al may leach out of the zeolite framework and form EFAl. This results in a decrease in Brønsted acid and an increase in Lewis acid sites, as seen for the commercial samples. However, for m-PL-ZSM-5, the amount of Brønsted acid sites remains unaltered, thus increasing the total amount of acid sites with more Lewis acid sites. This indicates that the additional secondary porosity in m-PL-ZSM-5 increases the accessibility of pyridine toward the Brønsted acid sites, which counteracts the decrease due to Al extraction.

2. Si/Al ratio and Al-concentration (ICP-OES), the number of Lewis (C_L) and Brønsted (C_B) acid sites (Py-FTIR) and the number of weak (C_w), medium (C_m) and strong (C_s) acid sites (NH₃-TPD).

			acid site concentrations [μmol g–1]
			Brønsted-Lewis			strength
catalyst	Si/Al	Al concentration [μmol g–1]	C L	C B	C tot,py	C w	C m	C s	C tot,NH3
c-ZSM-5	25.0	641	24	289	313	136	52	281	468
m-ZSM-5	24.0	666	46	226	272	118	63	226	408
PL-ZSM-5	29.2	551	22	118	140	55	53	152	260
m-PL-ZSM-5	28.6	563	31	118	149	75	81	170	326

The NH₃-TPD results are shown in Table (and their deconvoluted profiles in Figure S4 of the Supporting Information) and support the Py-FTIR results. Hierarchization of c-ZSM-5 results in a decrease in the total amount of acid sites, which is in accordance with literature and indicates the loss of some framework Al. , As observed with Py-FTIR, modification of PL-ZSM-5 leads to an increase in the total amount of acid sites accessible to NH₃, contrary to what is expected from both literature and the m-ZSM-5 results. The amount of acid sites accessible for NH₃ adsorption is also significantly lower for the platelike materials than for the conventional pair. This reduction of accessible acid sites is consistent with observations by Qureshi et al. and Li et al. in their platelike materials. Vieira et al. reported a reduced acid site accessibility in small ZSM-5 crystals compared to larger crystals with a similar Si/Al ratio and ascribed it to the presence of structural defects and/or amorphous domains. Interestingly, the desilication of PL-ZSM-5 results in an increased relative crystallinity, suggesting that some amorphous silica domains may be present in the parent PL-ZSM-5. This could lead to an overestimation of catalyst mass, and the removal of these amorphous domains during desilication could lead to an apparent increase in acid site concentration and mesoporosity per gram of catalyst. However, this hypothesis is not supported by the other characterization results. For instance, the XRD pattern of PL-ZSM-5 (Figure c) does not show an elevated background, which would be expected in the presence of significant amorphous content. Similarly, while SEM and TEM images (Figure e,f and Figure S1) reveal small particles in PL-ZSM-5 that could be amorphous, these domains persist after desilication (Figure g,h), suggesting they are not removed. FTIR analysis of the OH-stretching region (Figure ) shows that the internal silanol nests, often associated with structural defects or amorphous regions, are slightly more prominent in PL-ZSM-5 than in c-ZSM-5. Upon desilication, these silanol nests decrease for both materials, but the relative reduction is more pronounced in c/m-ZSM-5 (67%) than in PL/m-PL-ZSM-5 (54%), further suggesting that the amorphous domain removal is not the dominant factor in the increased acid site concentration. Alternatively, the increase in measured acid sites in m-PL-ZSM-5 could be due to the exposure of previously inaccessible framework Al upon desilication. In PL-ZSM-5, some Al atoms may be located in catalytically inaccessible T-sites, and desilication may enhance the accessibility of these sites. This is further supported by the relative acid site concentrations determined by each characterization method, i.e., the ratio of acid sites in PL-ZSM-5 to those in c-ZSM-5. Since NH₃ is a smaller probe molecule than pyridine, it can access more confined or partially obstructed sites. Accordingly, PL-ZSM-5 contains 56% of the acid sites of c-ZSM-5 when measured by NH₃-TPD, but only 44% when measured by Py-FTIR, indicating that a larger portion of the acid sites in PL-ZSM-5 are located in less accessible environments.

The change in EFAl is further assessed through ²⁷Al NMR, as shown in Figure . All four catalysts exhibit the characteristic intense peak with a maximum located around 55 ppm, attributed to the framework tetrahedral Al. ,− The minor differences in the shape of the peak have been attributed to the different distribution of Al atoms over the T-sites in the MFI structure. ,, The peak can be deconvoluted into different peaks with maxima between 45–50 ppm, 50–57 ppm and 57–65 ppm, as shown in Figure S5, each corresponding to one or more T-sites. It is suggested that the peaks at lower chemical shifts, with maxima between 45–50 ppm and 50–57 ppm, are related to the aluminum located in the channel intersections, whereas peaks with higher chemical shifts, with maxima between 57–65 ppm, are related to aluminum in the straight or sinusoidal channels. ,− When comparing the c-ZSM-5 and PL-ZSM-5, the platelike material’s peak maximum is located at 56 ppm compared to 55 ppm for c-ZSM-5. Furthermore, the band of PL-ZSM-5 extends more toward the higher chemical shifts, which can be deconvoluted into a peak with a maximum located at 62 ppm (Figure S5). This could indicate that most of the aluminum is situated inside the channels instead of in the intersections. After modifications, the peak maximum moves toward the lower chemical shifts (dashed lines in Figure ). Furthermore, both hierarchical zeolites show a broad signal ranging from 20 to 45 ppm, which has been attributed to Al clusters in a less-organized extra-framework phase and penta-coordinated Al ,, and is more significant in m-ZSM-5 compared to m-PL-ZSM-5. The presence of octahedrally coordinated EFAl is visible with a small peak at 0 ppm. ,− Both c-ZSM-5 and PL-ZSM-5 contain this peak, but it is much more significant in the platelike material. This corroborates the higher relative amount of Lewis acid sites which was observed in Py-FTIR. After modification, this EFAl peak almost completely disappears for the conventional pair, whereas for the platelike material it increases. This octahedrally coordinated EFAl may transform into clusters or penta-coordinated Al, causing the broad band between 45 and 20 ppm in m-ZSM-5. After modifying the platelike morphology, there is also an increase in the peaks associated with EFAl, however this is mainly the octahedrally coordinated Al. This indicates that the type of EFAl formed during modification slightly differs depending on the morphology of the parent H-ZSM-5. This could be due to the ease with which the etched Si and Al can diffuse out of the zeolite framework.

5.

²⁷Al NMR for c-ZSM-5 (red), m-ZSM-5 (dashed red), PL-ZSM-5 (blue), and m-PL-ZSM-5 (dashed blue). To account for the differences in material density, all peaks have been normalized to the peak at 55 ppm.

These characterization results show the effect of the parent morphology on the formed mesopores. When the parent zeolite is commercial H-ZSM-5, i.e., irregularly shaped crystals that tend to aggregate, small intracrystalline mesopores are formed, connecting the micropores of the zeolite. The TPA⁺-cations in the etching mixture tend to be more stable on the catalytic surfaces compared to Na⁺-cations, and combined with their steric hindrance, they protect the surrounding Si-atoms, preventing strong desilication and generating smaller mesopores which are uniformly spread throughout the crystals. ,,, When modifying platelike H-ZSM-5, the TPA⁺-cations seem to lose their protective character, resulting in large intracrystalline mesopores that sometimes merge into large macropores. As the Na⁺ can diffuse into the zeolite channels, but the TPA⁺ cannot, it is possible that while TPA⁺ protects the outer surface, NaOH diffuses into the micropores, initiating unprotected etching within the crystals. Perhaps the short b-axis makes it significantly easier for these Na⁺ cations to diffuse to the center of the crystals compared to the much longer diffusion path length in c-ZSM-5. This could create an attack both on the zeolite surface and inside the micropores, increasing the severity of the hierarchization and resulting in the large macropores found in m-PL-ZSM-5. Results of ICP-OES and pyridine FTIR (Table ) show a slight decrease in Si/Al ratio after modification of both morphologies. This decrease is rather small compared to typical desilication methods with solely NaOH. ,, As earlier reported in literature, TPAOH has a less selective character toward silicon extraction, thus losing the treatment’s selective character to Si and leaching more Al and preserving the Si/Al ratio of the parent zeolite to a greater extent. ,

3.2. Catalytic Performance

3.2.1. The Effect of the Morphology

Activity analysis is obtained at 483 and 513 K, and results are shown in Figure where the butanol conversion is plotted as a function of the site time. The site time, i.e., the number of BAS divided by the molar flow of the butanol feed, is used to compensate for the different number of BAS per gram of catalyst. The left graph compares c-ZSM-5 (filled symbols) to m-ZSM-5 (open symbols) with 483 K the light-colored symbols at the bottom, and 513 K the darker-colored symbols at the top of the graph. Similarly for the right graph, where PL-ZSM-5 (filled) and m-PL-ZSM-5 (open) are compared. In Figure S6 of the Supporting Information c-ZSM-5 and PL-ZSM-5 are shown in the same graph. At both temperatures, PL-ZSM-5 exhibits a higher conversion per active site compared to c-ZSM-5, although much more pronounced at 513 K than at 483 K. Furthermore, temperature screening of the catalysts was performed from 453 to 513 K and the turnover frequency (TOF) is plotted as a function of the temperature in Figure S7. The activity of PL-ZSM-5 is more temperature dependent than that of c-ZSM-5, resulting in a higher apparent activation energy for the platelike variant. Until 473 K, PL-ZSM-5 is less active, and from 473 K it surpasses c-ZSM-5. As the difference in activity increases with increasing temperature, the improved activity of PL-ZSM-5 is much more pronounced at 513 K compared to 483 K.

6.

Butanol conversion (X_BuOH) as a function of the site time for c-ZSM-5­(red filled circle), m-ZSM-5 (red open circle), PL-ZSM-5 (blue filled triangle), and m-PL-ZSM-5 (blue open triangle) at p _tot = 5 bar, p_BuOH = 29 kPa and temperatures 483 K (light) and 513 K (dark). Error bars indicate the 95% confidence interval.

Since the apparent activation energy differs between platelike and commercial ZSM-5, it cannot be excluded that the dominant reaction mechanism is different for the platelike and commercial ZSM-5. A higher apparent activation energy has been related to an increased contribution of the direct reaction paths compared to the etherification to DBE. ,, A higher energy barrier for the formation of DBE in PL-ZSM-5 could increase the catalyst’s activity since adsorbed dibutyl ether (DBE*) is relatively stable and can occupy over 90% of the acid sites, thereby “poisoning” the surface and inhibiting other (direct) reaction pathways. , However, an increased activity can also occur without selectivity changes due to a difference in the active site occupation and accessibility effects by, for example, the reduced path length in the straight channels, increasing the transport to and from the acid sites, and the increased amount of pore mouths to those straight channels. ,,,

As observed in the FTIR and NH₃-TPD results, the quantity and type of acid sites of PL-ZSM-5 and c-ZSM-5 differ, which could also affect their activity. Given that the activity is expressed as a function of site time, the effect of the quantity of acid sites is eliminated, indicating that the type of acid sites, and their accessibility, might influence the activity. It has been reported that for the n-butanol dehydration, a higher acidity, and more specifically more strong acid sites, results in an increased activity per acid site. , However, here the opposite occurs as the platelike ZSM-5 with the lowest amount of acid sites exhibits the highest activity per acid site. This leads us to hypothesize that the platelike morphology significantly affects the catalyst’s activity. To gain further insights, the selectivity-vs-conversion profiles are assessed and shown in Figure (and Figure S8 of the Supporting Information, where c-ZSM-5 and PL-ZSM-5 are shown in the same graph).

7.

Selectivity as a function of conversion for c-ZSM-5­(red filled circle), m-ZSM-5 (red open circle), PL-ZSM-5 (blue filled triangle), and m-PL-ZSM-5 (blue open triangle) at p _tot = 5 bar, p_BuOH = 29 kPa and temperatures 483 K (light) and 513 K (dark). Error bars indicate the 95% confidence interval.

At both temperatures, the selectivity toward isobutene is negligible, and all samples exhibit a relatively stable selectivity to 1-butene of about 0.2 mol mol^–1, corresponding well with the literature. − ,− Small differences in selectivity toward the 2-butenes and dibutyl ether are observed, i.e., PL-ZSM-5 exhibits a slightly higher selectivity toward the 2-butenes at the expense of dibutyl ether. N₂ sorption results showed the presence of a small amount of mesopores in c-ZSM-5, allowing larger transition states and products, such as dibutyl ether, to form more easily due to a reduced steric hindrance. Furthermore, the bulkier DBE is probably less likely formed inside the microporous zeolite channels due to confinement effects, , and since the ²⁷Al NMR results suggest that more Al is located in the straight and sinusoidal channels in PL-ZSM-5, the direct formation of butenes could be promoted in PL-ZSM-5. Consequently, decreasing the selectivity toward DBE and increasing the importance of the direct reaction mechanisms in the overall reaction rate. As these direct reaction mechanisms become more important at temperatures above 500 K, where the relative coverage of dibutyl ether on the catalyst’s surface reduces, ,, this smaller confinement in PL-ZSM-5 has more effect on the activity of the catalyst. Nevertheless, these selectivity differences are rather small, and other factors, such as the earlier mentioned fast transport toward and from the acid sites and the higher number of pore mouths can also affect the activity.

The zeolites’ stability is assessed through time on stream experiments performed at 513 K for 60 h. Results are shown in Figure and Figure S9, and reveal an increased stability for the platelike morphology, i.e., a relative decrease in the conversion of 6% for PL-ZSM-5 compared to 14% for c-ZSM-5 over 60 h. An increased amount of pore mouths and reduced path length have already been linked to an increased activity and stability for H-ZSM-5 in the ethanol-to-hydrocarbons reaction and n-butanol dehydration. The coke formation resulting in pore blockage has a smaller effect on active site accessibility compared to acid sites in zeolites with longer channels. Consequently, more coke formation can occur before a significant decrease in activity is observed, increasing the coke tolerance. Furthermore, the short b-axis will also improve the diffusion of the formed products and coke precursors out of the zeolite crystals, limiting secondary reactions that cause coke formation and accumulation of the formed cokes. ,,

8.

Butanol conversion as a function of time on stream (TOS) for c-ZSM-5­(red filled circle), m-ZSM-5 (red open circle), PL-ZSM-5 (blue filled triangle), and m-PL-ZSM-5 (blue open triangle) at p _tot = 5 bar, p_BuOH = 29 kPa and 513 K. The relative loss in conversion is indicated by the full lines for the parent zeolites and the dashed lines for the hierarchical zeolites.

3.2.2. The Effect of Mesopores in Commercial and Platelike H-ZSM-5

The performance of the two hierarchical zeolites is assessed through activity analysis at 483 and 513 K, as shown in Figure . The small mesopores of m-ZSM-5 positively affect the catalyst’s activity per active site at both 483 and 513 K. When comparing the apparent activation energies, obtained from the temperature screening measurements shown in Figure S6, no differences are observed. This suggests that the dominant reaction mechanism is unaltered after etching. The improved activity of the hierarchical zeolites is therefore linked to an increased accessibility of the active centers, allowing a more efficient catalytic cycle. Furthermore, the mesopores likely facilitate the desorption of dibutyl ether from the zeolite surface due to the reduced confinement within the zeolite. This minimizes the DBE-inhibition effect and increases the activity. The selectivity results (Figure ) support this assumption, i.e., an increase in DBE selectivity, at the expense of 2-butenes, is noticeable for m-ZSM-5 at 483 K compared to c-ZSM-5. As an increased coverage of the catalyst surface with dibutyl ether, due to an increased DBE-formation, would lead to a decrease in activity, , it is assumed that facilitated desorption of dibutyl ether causes the higher selectivity toward dibutyl ether. In addition, the mesopores could improve the diffusion of all reactants and products toward and away from the active sites and expose acid sites that were inaccessible before modification, ,,, increasing the overall activity without affecting the selectivity. Moreover, internal silanols and silanol nests are removed during the modification of the zeolites. These have been known to trap molecules that can block acid sites.

For m-PL-ZSM-5, the large mesopores only slightly improve the activity at 483 K and have no effect at 513 K (Figure ). Furthermore, a small increase in selectivity toward DBE at the expense of selectivity toward 2-butenes is observed at both temperatures (Figure ). It is assumed that two effects are occurring simultaneously. First, the mesopores improve the desorption of dibutyl ether, as demonstrated by its increased selectivity. As mentioned before, this has been shown to positively affect the activity of the catalyst. On the other hand, the large mesopores result in a decreased microporosity of m-PL-ZSM-5 compared to PL-ZSM-5 (Table ), which has a negative impact on the activity, as the confined environment and interactions with the framework stabilize transition states, leading to efficient catalysis. Since the introduction of secondary porosity in the platelike H-ZSM-5 barely affects the conversion per active site, it indicates the importance of the zeolite microporosity for the butanol dehydration reaction. This could also explain the high activity of PL-ZSM-5, with its presumably higher amount of active sites in the channels. The introduction of mesopores reduces microporosity in the zeolite crystal, and as mentioned before, the ether-mediated pathways may be less important in the platelike materials, thus reducing the positive effect of the enhanced DBE desorption.

Stability tests performed over 60 h of time on stream are shown in Figure and indicate that the mesopores in m-ZSM-5 have a negative effect on the catalyst’s stability as they increase the relative loss of conversion over 60 h from 14% for c-ZSM-5 to 23% for m-ZSM-5. This is contrary to what was expected, i.e., that the mesopores would improve the removal of coke precursors out of the catalyst, thus decreasing their oligomerization to cokes. This could indicate that the diffusion of coke precursors is not the most important factor in the deactivation of H-ZSM-5. When looking more closely at the type of mesopores formed in m-ZSM-5, they connect the existing micropores, but do not reduce the average diffusion distance from the surface to the core. Thus, when one of the micropores in the crystal is blocked close to the surface, all active sites further down the diffusion pathway are no longer accessible. However, this could indicate a similar deactivation rate as c-ZSM-5 but does not explain the reduced stability of m-ZSM-5. The FTIR results showed a reduced amount of internal silanols, which can trap coke precursors, but it also showed an increase in the number of Lewis acid sites. The ²⁷Al MAS NMR results showed that these Lewis acid sites are mainly penta-coordinated Al and distorted Al, which could enhance the coke formation and the mesopores could facilitate their growth as they provide extra space within the channels.

On the other hand, hierarchical platelike H-ZSM-5 shows a further improvement of the catalyst’s stability compared to the parent material, i.e., a relative decrease in the conversion of 3% compared to 6% over 60 h. The mesopores of m-PL-ZSM-5 are, however, much larger than those of m-ZSM-5. Hence, substantially more coke can be formed before pore blocking, further increasing the stability. Visual observations of the catalyst after the reaction (Figure S10) show that the used platelike zeolites have a gray color, whereas the color of the used conventional samples is rather yellow-brown, which might be due to a different type of coke formed. The hierarchical zeolites had a darker color for both sets, implying a larger amount of coke present. Due to catalyst pretreatment and setup limitations, characterization of these used catalysts is not possible, therefore, no firm conclusions on the type of coke can be made.

4. Conclusions

In this work, platelike, hierarchical, and hierarchical platelike H-ZSM-5 are created and compared to a commercial H-ZSM-5. The platelike zeolites significantly outperform the conventional ones in conversion per acid site, selectivity toward the butenes, and stability. The platelike morphology with its short b-axis has a reduced path length for the straight channels, which enhances the transport toward and away from the active sites. Consequently, it increases the catalyst’s activity and stability. Furthermore, the crystals have a large smooth ac-plane, providing a large amount of pore mouths to the straight channels, further promoting activity and stability. The NMR results indicated that the platelike H-ZSM-5 presumably contains relatively more active sites inside the microporous channels. This could enhance the participation of direct reaction mechanisms, as confinement effects in these microporous channels promote the formation of smaller molecules. By reducing the contribution of the ether-mediated reaction paths, inhibition effects of dibutyl ether are suppressed, and the activity and stability of the catalyst are improved, accompanied by a reduced dibutyl ether selectivity.

The second strategy of improving the catalyst by adding secondary porosity through base-leaching improves the catalyst’s performance, although less pronounced than by changing the zeolite’s morphology. The etching mixture of NaOH and TPAOH has proven to result in a mild etching procedure, creating small intracrystalline mesopores without losing the original microporosity of the commercial ZSM-5. These mesopores improve the accessibility of the active sites and enhance the dibutyl ether desorption of the zeolite surface, leading to an improved activity per acid site. Nonetheless, these small mesopores do not improve the stability, as the hierarchical H-ZSM-5 showed the highest deactivation over time. This is ascribed to the formation of additional Lewis acid sites upon modification, which has been known to enhance the deactivation rate in zeolites.

The same etching procedure was used for platelike ZSM-5, but resulted in a different type of mesopores, and large mesopores and sometimes even macropores were formed. Despite the formation of these large secondary pores, the effect on the catalytic performance is limited, only slightly improving the activity and stability of the platelike H-ZSM-5. Since the platelike morphology already considerably optimizes the catalyst’s performance, the mesopores only have a limited effect. Furthermore, as it is assumed that the indirect reaction pathways are less important in the platelike zeolite, an enhancement in the dibutyl ether desorption has a smaller impact on the activity.

These results imply the importance of the microporous confinement for the n-butanol dehydration, as the improved accessibility to the microporous channels in the platelike H-ZSM-5 showed the best results. Moreover, as the addition of mesopores did not show a significant effect, it is suspected that these mesopores can only create molecular highways, improving the transport to and from the acid sites, but cannot change the dominant reaction path for the butanol dehydration.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsengineeringau.5c00033.

• SEM images of m-ZSM-5 and m-PL-ZSM-5; combined N₂ sorption isotherms of c-ZSM-5 and PL-ZSM-5; BJH pore size distributions; deconvoluted NH₃-TPD profiles; deconvoluted ²⁷Al MAS NMR profiles; butanol conversion as a function of site time for c-ZSM-5 and PL-ZSM-5; TOF as a function of temperature, selectivity as a function of butanol conversion for c-ZSM-5 and PL-ZSM-5; Butanol conversion as a function of time on stream for c-ZSM-5 and PL-ZSM-5; visual observations of catalysts after reaction (PDF)

CRediT: Phebe Lemaire data curation, formal analysis, investigation, validation, visualization, writing - original draft, writing - review & editing; Arno de Reviere conceptualization, investigation, methodology, writing - original draft, writing - review & editing; Dhanjay Sharma methodology, writing - review & editing; Valerie Ruaux formal analysis, investigation, methodology; Jaouad Al Atrach formal analysis, investigation, methodology; Valentin Valtchev project administration, resources, writing - review & editing; Joris W. Thybaut funding acquisition, resources, supervision, writing - review & editing; Maarten K. Sabbe conceptualization, funding acquisition, project administration, supervision, writing - review & editing; An Verberckmoes conceptualization, funding acquisition, project administration, resources, supervision, writing - review & editing.

The authors declare no competing financial interest.