One-Pot Direct Synthesis of b-Axis-Oriented and Al-Rich ZSM-5 Catalyst via NH₄NO₃-Mediated Crystallization for CO₂ Hydrogenation

Abstract

Al-rich NH₄-ZSM-5 with highly oriented crystals was directly synthesized through a one-pot hydrothermal technique, using ammonium nitrate as a metal-free mineralizer. The samples were characterized by XRD, N₂ adsorption–desorption, SEM, FTIR, Py-FTIR, ²⁷Al MAS NMR, ²⁹Si MAS NMR, ¹H MAS NMR, and TGA techniques. The impact of aluminum source, ammonium source, and H₂O/SiO₂ molar ratio was studied. XRD results showed that the ZSM-5 catalyst with a low Si/Al ratio (13) was successfully synthesized without any amorphous phase, including a microporous/mesoporous structure. A low H₂O/SiO₂ molar ratio (75) resulted in coffin-shape surface morphology, large b-axis-oriented particles (ca. 19 µm), and high specific surface area (>300 m² g⁻¹), providing a large portion of straight channels (90.5%). The catalytic activity of the catalysts was evaluated in the CO₂ hydrogenation reaction in tandem configuration with a Na/Fe₂O₃ catalyst. The results confirmed that highly b-oriented crystals improved the product shape selectivity to p-xylene by affecting the diffusion resistance. Therefore, the developed catalyst provided high CO₂ conversion (45%) and high aromatic selectivity (77%), with p-xylene accounting for 82% of the produced xylene compounds, over a long-term time on stream (17 h). These results demonstrate the effectiveness of the direct synthesis strategy in producing Al-rich ZSM-5 catalysts with tailored textural and acidic properties for tandem and shape-selective catalysis.

1. Introduction

Zeolites are attractive materials as efficient catalysts or adsorbents owing to their unique textural/acidity properties and high chemical/thermal stability. ZSM-5 is one of the most common types of zeolites, with a pore diameter of 0.54 nm and a wide range of Si/Al ratios. ZSM-5 catalysts are being applied in numerous processes, including methanol to olefins [1], furan deoxygenation [2], biodiesel production [3], and CO₂ hydrogenation [4] to name just a few. It is worth noting that alkali ions, especially Na⁺, are used in the conventional synthesis of ZSM-5, while the H-form of ZSM-5, which can be prepared through ion-exchange process of Na-ZSM-5, is the one active as an acid catalyst. In general, at least one ion-exchange process is required to exchange Na⁺ ions with NH₄⁺ ions to form H-ZSM-5 through the decomposition of NH₄⁺ ions by calcination at a high temperature. These processes can lead to some water pollution and high energy consumption. Direct synthesis of H-ZSM-5 using ammonium ions instead of alkali ions can eliminate the ion-exchange step, leading to a faster and more repeatable synthesis process. Ammonium hydroxide (NH₄OH) and tetrapropyl ammonium hydroxide (TPAOH) are the most common ammonium ion source and structure directing agent (SDA), respectively, for the direct NH₄-ZSM-5 synthesis [5]. Bibby et al. [6] reported the direct synthesis of NH₄-ZSM-5 using NH₄OH for the first time. Wu et al. [7] applied NH₄F through a solvent-free route for direct synthesis of NH₄-ZSM-5. Feng et al. [8] used NH₄OH and glucose, the latter as an additive, to directly prepare NH₄-ZSM-5 (Si/Al = 25). Xue et al. [5] studied the direct synthesis of nano-sized NH₄-ZSM-5 with a high Si/Al ratio (>15) using ammonium solution and TPAOH template in a silicalite-1 seed-induced technique. They found that samples with Si/Al ratios lower than 30 were amorphous silica without any ZSM-5 crystalline phase. A similar result was reported by Hou et al. [9]. This can be explained by the reduction in nucleation and growth rates at a high aluminum (Al) content [5]. It is reported that Al sitting is significantly controlled by the surrounding electronic environment because Al species should be located near the charge balance agent [10]. Within the ZSM-5 framework, there are 12 different tetrahedral sites (T sites) that are located at straight or sinusoidal channels or channel intersections. The synthesis conditions are mainly effective for controlling the Al distribution through T sites in the ZSM-5 framework [11]. Precisely controlling the location of the Al species offers a promising approach to optimize the acidity as well as the catalytic properties of the ZSM-5 catalyst. It is well documented that more Al atoms at the channel intersection favor aromatization reactions, while Al atoms at channels enhance olefin formation in methanol to olefins [4]. The 3D pore structure of ZSM-5 includes a sinusoidal channel (0.51 × 0.55 nm) and a straight channel (0.53 × 0.56 nm), which are along the a-axis and b-axis crystallographic directions, respectively. Therefore, crystal orientation can influence the percentage of straight and sinusoidal channels as well as the catalytic activity. Several research works have shown that the sinusoidal channels, involving higher diffusion resistance, accelerate the catalyst deactivation while the straight channels, with lower diffusion resistance, favor a long-term catalytic lifetime [12]. Therefore, a b-axis-oriented ZSM-5 catalyst can promote product selectivity in tandem reactions owing to the improved shape selectivity of the straight channels as well as the appropriate residence time of intermediate compounds to form desired products. For instance, Wang et al. [13] maximized sinusoidal channels in micro-sized ZSM-5 particles to enhance xylene isomerization in the methanol/toluene alkylation reaction, leading to high p-xylene selectivity.

Zhao et al. [14] studied ZSM-5 synthesis with different anionic species derived from varied Al sources. They found that the presence of NO₃⁻ in the synthesis gel influenced the electron environment and pathway of the framework formation, enhancing the incorporation of Al species during the early stage of ZSM-5 crystallization. They reported that aluminum nitrate (Al(NO₃)₃) as the Al source resulted in particles with a-axis, c-axis, and b-axis lengths of 100 nm, 880 nm, and 370 nm, respectively. Guo et al. [15] studied the impact of different alkali metal cations and their counter anions on the ZSM-5 synthesis. The results showed different crystallization pathways for the Na⁺ and K⁺ systems, leading to different crystal morphology. However, the role of anions in the ZSM-5 crystallization and the Al sitting needs to be studied thoroughly. More information on these effects is essential to optimize the ZSM-5 synthesis and the resulting catalytic properties.

To the best of our knowledge, there is no report on using ammonium nitrate as a metal-free mineralizer in the direct synthesis of NH₄-ZSM-5 with a very low Si/Al ratio (<13). In the present work, a one-pot hydrothermal technique was applied to synthesize b-oriented Al-rich NH₄-ZSM-5 with a large particle size along the b-axis and c-axis, leading to a high-volume fraction of the straight channels as well as longer sinusoidal channels. Furthermore, the catalytic activity of the catalysts was evaluated in the CO₂ hydrogenation reaction to achieve high aromatic and p-xylene selectivity for a long-term time on stream.

2. Materials and Methods

2.1. Materials

The precursors for the catalyst synthesis, namely sodium aluminate (NaAlO₂, 55 wt.% Al₂O₃), aluminum isotropies (Al(O-i-Pr)₃, >98%), tetrapropyl ammonium bromide (TPABr, C₁₂H₂₈BrN, >99 wt.%), ammonium nitrate (NH₄NO₃, 99 wt.%), sodium hydroxide (NaOH, 99.6 wt.%), tetrapropyl ammonium hydroxide (TPAOH, C₁₂H₂₉NO, 40 wt.% in H₂O), ferrous sulphate heptahydrate (FeSO₄·7H₂O), sodium carbonate (Na₂CO₃), and tetraethyl orthosilicate (TEOS, 99%) were obtained from Sigma Aldrich (Saint Louis, MO, USA) and Fisher Scientific (Waltham, MA, USA).

2.2. Catalyst Preparation

2.2.1. Metal Catalyst

The Fe₂O₃ catalyst was prepared using the precipitation method. Typically, 7.0 g of FeSO₄·7H₂O was dissolved in 20 mL of H₂O to obtain the solution A (1 M). Then, 10.2 g of Na₂CO₃ was dissolved in 100 mL of water (solution B). Solution A was added dropwise into solution B at 50 °C under magnetic stirring for 1 h. The obtained solid product was filtered, washed three times with water, dried, and calcined at 550 °C for 6 h with a ramp rate of 2 °C min⁻¹. A calculated amount of Na₂CO₃ was added to the Fe₂O₃ catalyst via the incipient wetness impregnation method (IWI) to obtain 4 wt.% of Na in the material, which was denoted as Na/Fe₂O₃.

2.2.2. ZSM-5 Catalyst

The Al-rich ZSM-5 catalysts (Si/Al = 13) were synthesized by the hydrothermal method. As the first step, a solution of ammonium nitrate, aluminum source, and distilled water was stirred for 15 min. TPABr was then added to the gel and stirred for 45 min (solution A). Simultaneously, TEOS was added to distilled water (solution B). Solution A was added to solution B dropwise under continuous agitation and stirred for 4 h. The pH of the solution was adjusted to 10.5 using TPAOH (less than 4 mL). The molar composition of the final solution was 3948H₂O:26.3SiO₂:1Al₂O₃:11.8template:xNH₄NO₃:yNa₂O. The catalysts were denoted based on the applied composition of the synthesis gel, as shown in

Table 1. Composition of the synthesis gels for the different catalysts investigated.

Catalyst	Al Source	Template	NH4+ Source	H2O/SiO2	x (NH4+/Al)	y (Na+/Al)
ZM-1	NaAlO2	TPABr	NH4NO3	150	1	1
ZM-2	NaAlO2	TPABr	NH4NO3	150	2	1
ZM-3	Al(O-i-Pr)3	TPABr	NH4NO3	150	1	0
ZM-4	NaAlO2	TPABr	-	150	0	2
ZM-5	NaAlO2	TPABr	-	150	0	1
ZM-6	NaAlO2	TPAOH	-	150	0	1
ZM-7	NaAlO2	TPABr	NH4OH	150	1	1
ZM-8	NaAlO2	TPABr	NH4NO3	75	1	1
ZM-9	NaAlO2	TPABr	NH4NO3	10	1	1

. It is worth noting that NaOH was not used for the synthesis of the catalysts with y = 1, and Na⁺ cations were only provided by NaAlO₂. For the ZM-7 and ZM-8 catalysts, the amount of water was reduced to obtain H₂O/SiO₂ ratios of 75 and 10, respectively. The crystallization was performed in a static stainless steel Teflon-lined autoclave at 180 °C under autogenous pressure for 48 h. After filtration and washing, the powder was dried at 110 °C for 12 h and calcined at 540 °C for 12 h (heating rate of 3 °C min⁻¹) in a muffle furnace under airflow.

2.3. Characterization Methods

Wide-angle X-ray diffraction (XRD) analyses were performed with a Siemens (Munich, Germany) 80 model D5000 diffractometer using CuKα radiation (λ = 0.15496 nm), operating at 40 kV and 40 mA. Data were collected in the 2θ range of 5–50° with a step size of 0.02° and a scanning rate of 2° min⁻¹. Textural properties were determined from the N₂ physisorption isotherm at 77 K obtained using a Quantachrome (Boynton Beach, FL, USA) Nova 2000 series instrument. Prior to measurement, samples were degassed under vacuum at 300 °C for 4 h to remove adsorbed moisture and impurities. The total surface area (S_BET) and total volume (V_meso) were determined by the Brunauer–Emmet–Teller (BET) isothermal equation within the p/po range of 0.05–0.2 and the pore diameter was estimated by density functional theory (DFT) mode. The t-plot method was applied to calculate the micropore volume (V_micro). The mesopore volume (V_meso) is the difference between the calculated total data and the corresponding micropore value. FT-IR measurements were performed in a Nexus Model infrared spectrophotometer (Nicolet Co., Mountain, WI, USA) at the resolution of 4 cm⁻¹. Self-supported wafer containing the powder in KBr was prepared. Scanning electron microscopy (SEM) was performed using a KYKY (Model, EM3200, Beijing, China) device at a potential difference of 26 kV. Prior to analysis, samples were dispersed on conductive carbon tape and sputter-coated with a thin layer of gold or platinum to minimize charging effects. X-ray fluorescence (XRF), Rigaku (Tokyo, Japan) ZSX Primus II, was employed to determine and quantify the elemental composition of as-prepared zeolites. Solid-state magic angle spinning (MAS) nuclear magnetic resonance (NMR) spectra were obtained with a Bruker (Billerica, MA, USA) Avance 300 MHz spectrometer. ²⁷Al MAS NMR spectra were collected at a resonance frequency of ~104 MHz, using a spinning rate of 10–12 kHz, a pulse width of ~1 µs, and a recycle delay of 1 s. ²⁹Si MAS NMR spectra were recorded at ~79 MHz, with a spinning rate of 5–8 kHz and a recycle delay of 30–60 s. Ammonia-temperature desorption (NH₃-TPD) was performed using an RXM-100 analyzer (Advanced Scientific Designs, Inc., Millersburg, OH, USA). Prior to each test, about 70 mg of sample was preheated at 500 °C for 1 h in an argon flow (30 mL min⁻¹), then cooled down to ambient temperature. The sample was saturated in NH₃ stream for 5 min at ambient temperature, followed by flushing in argon flow (50 mL min⁻¹) for at least 12 h. A diluted solution of bromothymol blue was employed to test complete evacuation of the physisorbed-NH₃ in the outlet stream. The analysis was carried out from room temperature to 650 °C at a ramp rate of 10 °C min⁻¹ under argon flow (30 mL min⁻¹). For FT-IR spectroscopy of chemisorbed pyridine (Py-FTIR), in situ diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) experiments using pyridine as a probe molecule were performed on a PerkinElmer spectrum 3 equipped with a liquid nitrogen-cooled Hg-Cd-Te (MCT) detector and a Praying Mantis cell from Harrick Scientific (Mount Kisco, NY, USA). A sample was loaded in the cell and in situ preheated at 300 °C for 1 h under vacuum to remove all moisture. The sample was then cooled down to 150 °C, followed by saturating in a pyridine/argon flow for 0.5 h and evacuation for 1 h. The IR spectra were acquired as an accumulation of 32 scans at a resolution of 4 cm⁻¹.

2.4. Experiments

CO₂ hydrogenation experiments were carried out in a fixed-bed reactor. The reactor was a stainless steel tube, including a K-type thermocouple probe near the catalyst bed. The bed included two separate zones where the Na/Fe₂O₃ metal catalyst (0.5 g) and ZSM-5 catalyst (0.5 g) were located in the top and bottom zone, respectively. The catalysts were fixed in the middle of the reactor using quartz wool on top, between, and at the bottom of the zones. The general operation conditions were 320 °C, 2.5 MPa, and GHSV of 4000 mL g⁻¹ h⁻¹. The feed was a gas mixture of CO₂ and H₂ (H₂/CO₂ = 3). Before the reaction, the catalyst was heated from 25 °C to 350 °C at a rate of 10 °C min⁻¹. Then, a flow of hydrogen was used to activate the catalyst at 350 °C for 1 h. During the reaction, the gas products were analyzed online using a GC HP-5890 chromatograph (Agilent Technologies, Inc., Santa Clara, USA) equipped with a thermal conductivity detector (TCD) and a HayeSep N packed column. The liquid products, obtained via condenser, were analyzed and quantified using GC-MS (Agilent 8890 GC, Santa Clara, CA, USA) equipped with a capillary column DB-WAX Ultra inert.

CO₂ conversion and product selectivity were calculated on a carbon atom basis according to the following equations:

$$Conversion,\%={\displaystyle \frac{{\left({CO}_2\right)}_{in}-{\left({CO}_2\right)}_{out}}{{\left({CO}_2\right)}_{in}}}\times 100$$

(1)

where ${\left({CO}_2\right)}_{in}$ and ${\left({CO}_2\right)}_{out}$ represent $\left({CO}_2\right)$ the molar flow rate at the inlet and the outlet of the reactor, respectively.

Product selectivity was calculated as shown in Equation (2):

$$S_i,\%={\displaystyle \frac{mol C_i\times i}{\sum _{i=1}^{n}mol C_i\times i}}\times 100$$

(2)

where C_i is an individual hydrocarbon product molar concentration and $\sum _{i=1}^{n}mol C_i\times i$ is the number of moles of carbon in the products.

STY (space–time yield) of products is calculated using Equation (3):

$$STYi={\displaystyle \frac{n_{Ci}}{m_{catalyst}\times t}}$$

(3)

where $n_{Ci}$ is the amount of mmol of product collected over time t (h). m_catalyst is the mass of catalyst (g).

The carbon balance was calculated as shown in Equation (4):

$$Carbon balance,\%={\displaystyle \frac{n{\left(C\right)}_{products}}{n{\left(C\right)}_{converted}}}\times 100$$

(4)

where n(C)_products is the number of moles of C atoms in the detected gas and liquid products and n(C)_converted is the number of moles of C atoms in the converted CO₂.

3. Results and Discussion

3.1. Characterization

Figure 1 shows the XRD patterns of the synthesized catalysts. The patterns of the ZM-1, ZM-2, ZM-3, and ZM-8 catalysts display the characteristic peaks related to the MFI framework (JCPDS: 44-0003 [16]). Increasing the amount of NH₄⁺ ions in the synthesis gel leads to lower peak intensities compared with the ZM-1 catalyst, owing to the low growth rate [17]. The lack of a broad feature centered at 23.5° indicates the non-existence of an amorphous silica phase for these catalysts. The synthesis of the ZM-1 catalyst has been repeated three times successfully, confirming the repeatability of the developed technique. The impact of Na impurity from NaAlO₂ on the direct synthesis of NH₄-ZSM-5 was studied by comparing catalysts ZM-1 and ZM-2 with ZM-3, which was synthesized using Al(O-i-Pr)₃ as the Al source. The ZSM-5 crystalline phase of the ZM-3 confirms the capability of the NH₄NO₃ to act as a metal-free mineralizer agent in the direct synthesis of Al-rich NH₄-ZSM-5 through a Na-free system. It is confirmed that the ionic Na impurity in the NaAlO₂, applied as the Al source, does not influence the crystallization owing to the very low Na concentration. Li et al. [18] reported that the ZSM-5 crystalline phase can only be obtained at a Na⁺/Si molar ratio of 0.6 for Al-rich ZSM-5 catalysts (Si/Al < 15). Five characteristic peaks of ZSM-5 catalysts at 2θ = 7.96°, 8.83°, 23.18°, 23.99°, and 24.45° were selected to determine the relative crystallinity (based on the ZM-3 catalyst), calculated by summing the peak area for each catalyst [8]. The results in Table 2 show that the ZM-3 catalyst has the highest relative crystallinity compared with the others. These results are in line with the literature [19].

Synthesis of the ZM-5 and ZM-6 catalysts was performed without NH₄NO₃ and NaOH using TPABr and TPAOH as templates, respectively. The XRD results show that these catalysts are amorphous silica, highlighting the role of NH₄NO₃ in the ZSM-5 synthesis. It is accepted that TPAOH is a stronger SDA than TPABr because it can act as a template and a mineralizer agent simultaneously. However, the high cost of TPAOH limits its application in industry, while TPABr is a cheap and cost-effective template for use at an industrial scale [20]. In order to compare NH₄⁺ and Na⁺ cations in the ZSM-5 crystallization, the ZM-4 sample was synthesized using NaOH at the same cation/Al molar ratio (Na⁺/Al = 1) as the ZM-1 catalyst (NH₄⁺/Al = 1). The XRD pattern of the ZM-4 includes a broad diffraction band at 22–27°, indicating an amorphous silica phase. Li et al. [18] reported that the ZSM-5 crystalline phase can only be obtained at a Na⁺/Si molar ratio of 0.6 for Al-rich ZSM-5 catalysts (Si/Al < 15). Karim et al. [21] reported that the Na content influenced the polymerization reaction of silicate as well as the substitution rate of Al to Si species in the framework. Therefore, the synthesis failure for ZM-4 can be explained by the low amount of Na⁺ (Na⁺/Al = 1). Yuan et al. [22] reported that different anions (Cl⁻, Br⁻, I⁻) in the synthesis gel led to the variation in ZSM-11 acid site location owing to the different covalent interaction with tetrabutylammonium (TBA⁺). This may affect the nucleation step rate as well as the location of Al species. Pashkova et al. [23] concluded that the type of anions in the synthesis gel influenced the distribution of Al pairs in the ZSM-5 framework. The anions with low polarizability favored a less positive charge on the nitrogen atom, leading to more Al atoms localized near the tetrapropyl ammonium (TPA⁺). Lario et al. [24] applied a density DFT study and found that the ionic radii of the cation (Na⁺ or K⁺) determined their capabilities for incorporation in the chabazite framework as well as Al distribution. They found that the use of K⁺ (ionic radii of 1.38 Å) allowed CHA zeolite (Si/Al = 10–16) crystallization from a synthesis solution with a lower amount of organic template compared with Na⁺ (ionic radii of 1.02 Å) owing to the higher structure-directing ability of K⁺ cation. The larger ionic radii of NH₄ (1.48 Å) can result in a behavior like K⁺ in terms of accelerating ZSM-5 crystallization.

The index peaks at 8.8° and 23° are assigned to [010] and [501] crystallographic faces of the ZSM-5 crystal, respectively, so that the ratio of their intensities can represent the volume fraction of b-oriented crystals. A large index peak at 8.8° confirms the synthesis of b-oriented crystals through the direct synthesis of NH₄-ZSM-5 using a NH₄NO₃ mineralizer. The ZM-1 catalyst displays the highest fraction of b-oriented crystals (3.57) compared with the ZM-2 (1.72), ZM-3 (2.30), and ZM-8 (2.3) catalysts.

Table 2. Framework data of the H-ZSM-5 catalysts.

Catalyst	Crystallite Size 1, nm	Relative Crystallinity, %	Straight Channel 2, %	Sinusoidal Channel 3, %	(Si/Al)bulk 4
ZM-1	0.95	84	79.59	20.41	8.2
ZM-2	1.58	72	84.82	15.18	5.0
ZM-3	1.65	100	87.35	12.65	7.6
ZM-8	1.59	85	90.50	9.50	8.0

¹ Determined using the Debye–Scherrer equation; ² determined by the method reported in [25]; ³ determined by (100 straight channel); ⁴ determined by XRF.

Table 2. Framework data of the H-ZSM-5 catalysts.

Catalyst	Crystallite Size 1, nm	Relative Crystallinity, %	Straight Channel 2, %	Sinusoidal Channel 3, %	(Si/Al)bulk 4
ZM-1	0.95	84	79.59	20.41	8.2
ZM-2	1.58	72	84.82	15.18	5.0
ZM-3	1.65	100	87.35	12.65	7.6
ZM-8	1.59	85	90.50	9.50	8.0

¹ Determined using the Debye–Scherrer equation; ² determined by the method reported in [25]; ³ determined by (100 straight channel); ⁴ determined by XRF.

During direct synthesis of NH₄-ZSM-5 catalysts, both TPA⁺ and NH₄⁺ cations can be incorporated into the lattice to compensate the negative charge in the ZSM-5 framework. TGA analysis can be used to characterize these cations in the NH₄-ZSM-5 catalysts with different amounts of NH₄⁺ cation in the synthesis gel (Figure 2a). The weight loss in the range of 50–200 °C is assigned to the evaporation of physically adsorbed water. The weight loss in the range of 200–450 °C indicates the combustion of organic template. There is another weight loss step in the range of 450–500 °C, which is more obvious in the DTG curve (Figure 2b). This step can be attributed to decomposition of the NH₄⁺ cation that interacts with the framework to compensate the negative charge [5]. The ZM-2 catalyst shows higher weight loss (9.8%) than the ZM-1 catalyst (9.4%) because of the higher NH₄⁺/Al (2) ratio in the synthesis gel.

As shown in Figure 3, the morphology of the catalysts is coffin-shaped particles at a micron scale. For the NH₄⁺ system, the interaction of these cations and the aluminosilicate anions is the main factor of crystallization owing to the large molecular size of NH₄⁺ and TPA⁺ cations. In addition, there is a weak interaction between NH₄⁺ and water molecules so that the “hydrated structure” value of NH₄⁺ (1.54 Å) is very much lower than that of the alkali cations [9]. As a result, less association of TPA⁺ and aluminosilicates leads to limited nucleation rate, leading to larger particle size. The ZM-1 catalyst exhibits the largest particle size compared with the ZM-2 and ZM-3 catalysts. The higher polarizability of the isopropoxide anion compared with the NO₃⁻ anion enhances the fluidity of the synthesis solution as well as the nucleation process. Therefore, the relative growth rate of the ZM-3 is low and smaller particles are formed. Zhao et al. [14] studied the effect of different aluminum sources on the crystallization of ZSM-5 particles. They found that the presence of the NO₃⁻ anion resulted in the largest particle size (880 nm) and the largest thickness of the b-axis (370 nm). The b-axis thickness of the ZM-1 and ZM-3 catalysts is 1.25 and 0.79 µm, respectively. It is well established that straight channels (along the b-axis) provide lower diffusion resistance compared with sinusoidal channels (along the c-axis) so that the high number of straight channels and longer sinusoidal channels of the ZSM-5 catalyst will bring exceptional advantages for catalytic reactions, reducing reactants and product diffusion resistance as well as coke formation. For instance, it is confirmed that, in CO₂ hydrogenation, large ZSM-5 crystals show higher shape selectivity of p-xylene than small crystals because the longer diffusion path promotes m- and o-xylene isomerization to p-xylene [26].

Figure 4 shows N₂ adsorption–desorption isotherms. The catalysts show type I (ZM-2 and ZM-8) and IV (ZM-1 and ZM-3) isotherms according to IUPAC classification, indicating micropores and mesoporous structures. The H4 type hysteresis indicates slit-shaped meso- and macropores in the catalysts [27]. The calculated textural data (

Table 3. Textural data of the H-ZSM-5 catalysts.

Catalyst	SBET (m2 g−1)	Sexternal (m2 g−1)	Vtotal (cc g−1)	Vmicro (cc g−1)	Vmeso (cc g−1)	DFT Pore (nm)	HF *
ZM-1	368	144	0.20	0.09	0.11	1.1	0.18
ZM-2	114	51	0.08	0.03	0.05	3.1	0.17
ZM-3	258	99	0.19	0.06	0.13	2.8	0.12
ZM-8	306	83	0.15	0.08	0.07	3.5	0.15

* Hierarchical factor = (S_meso/S_BET) × (V_micro/V_total).

) represent the higher surface area (368 m² g⁻¹) and total pore volume (0.2 cc g⁻¹) of the ZM-1 catalyst compared with the others. The high external surface area of the ZM-1 catalyst (144 m² g⁻¹) can result from intracrystalline mesopores [28]. Increasing the content of NH₄NO₃ in the synthesis gel (ZM-2) decreases the particle size compared with the ZM-1 catalyst owing to the reduced growth rate [17]. ZM-1 has the highest HF factor (0.18), which indicates its more hierarchical structure compared with the other catalysts. The average pore size of the catalysts reflects their mesoporous structure.

Figure 5 shows the FTIR spectra of the catalysts. The bands at 469, 800, and 1100–1200 cm⁻¹ are assigned to the symmetric stretching vibration of TO₄ (T = Si or Al) bonds [29,30]. The band at ca. 545 cm⁻¹ is attributed to the asymmetric tension of the five-membered rings of the ZSM-5 framework. The band at 1456 cm⁻¹ is attributed to N-H bending vibration modes of the NH₄⁺ moieties, which is in line with the band at 1887 cm⁻¹, resulting from the adsorption of NO/NO_x compounds (produced through NH₄⁺ decomposition) [31]. The band at 1639 cm⁻¹ is associated with OH bending vibrations and physically adsorbed water [32,33]. The band at ca. 2925 cm⁻¹ is assigned to the C–H stretching of the TPA⁺ template [34]. The presence of hydroxyl groups is indicated by the band at 3445 cm⁻¹ [30,35].

Figure 6 shows ²⁷Al MAS NMR results for the ZM-1 and ZM-3 catalysts to study the impact of aluminum source in Al siting through the direct synthesis of ZSM-5. There are two distinct peaks at ca. 53 ppm and ca. 0 ppm, which are assigned to four-coordinated framework and six-coordinated extra-framework Al (EFAl) species, respectively [36]. In the spectrum of ZM-1, the peak at 53 ppm is stronger than the peak at 0 ppm, indicating that the majority of Al species are in tetrahedral coordination and incorporated in the ZSM-5 framework. The ZM-3 catalyst includes more extra-framework. These results can be related to the smaller crystal size of ZM-3 and the higher content of EFAl species. The peak located at ca. 53 ppm can be deconvoluted into several peaks, including maximums at 52, 53, 54, 56, and 58 ppm. It is reported that the peak with a maximum at 56 characterizes the Al species located in the channel intersections, whereas the peak with a maximum at 54 ppm is attributed to the Al located in the straight/sinusoidal channels [37]. The deconvolution results in

Table 4. Al species distribution in the H-ZSM-5 catalysts.

Catalyst	Chemical Shift (ppm) Assignment of Al Sites and Relative Peak Areas, %					56 + 54, %	54/56	(Si/Al)framework *
	52	53	54	56	58
ZM-1	0.1	1.4	33.3	63.0	2.0	96.3	0.5	12.8
ZM-3	1.0	39.5	8.2	8.6	42.5	16.8	0.9	20.4

* Based on ²⁹Si MAS NMR.

show that the ZM-1 sample comprises more Al species (63%) in the channel intersections compared with ZM-3 (8.6%). This is expected to be beneficial for aromatic production through CO₂ hydrogenation reaction. The larger size of the intersection zone (0.64 nm) compared with the channel zone (ca. 0.54 nm) is believed to provide sufficient space for aromatization reactions [4]. Furthermore, the ZM-3 shows a clear broad signal at ca. 30 ppm, related to Al clusters with less-organized extra-framework phase and penta-coordinated Al species [38]. The ZM-3 catalyst includes more EFAl in agreement with the XRF results, which indicate the lower bulk Si/Al ratio of the ZM-3 sample (7.6) compared with ZM-1 (8.2) (Table 2).

The location, concentration, and strength of Brønsted acid sites significantly impact the catalytic performance of ZSM-5 catalysts. The reactivity of these active sites is determined by the concentration, spatial arrangement, and local environment of Al species in the framework. DFT simulation results confirmed that kinetic factors are more important for Al siting during the synthesis compared with thermodynamic factors [39]. The high content of framework Al species implies a high concentration of Brønsted acid sites at the channel intersection. The strength of the Brønsted acid sites can be tuned through hydrothermal/calcination treatments by dealumination of the framework and formation of EFAl species [40]. It is worth noting that EFAl species involve Lewis acid sites, leading to improved catalytic activity. Almutairi et al. [41] found that the FAU-type zeolite catalyst, including EFAl species, showed higher performance in the catalytic cracking of propane due to the synergistic effect of the Brønsted acid sites and EFAl species. In fact, the existence of bulkier EFAl species in zeolite micropores can change the zeolite confinement space, which can enhance the non-covalent stabilization of intermediate compounds throughout the catalytic reaction [42].

The ²⁹Si MAS NMR spectrum of the catalysts includes a dominant resonance signal at ca. −113 ppm that is assigned to the Si in SiO₄ tetrahedron units (Figure 7). A careful deconvolution of the spectra shows four peaks. The peak at −103 ppm is related to the terminal silanol groups [43]. The peaks at −113.7 and −113.4 ppm are attributed to resonances of the Si surrounded by 4Si (4Si, 0Al) units. The peaks at −107 and −107.3 ppm stem from resonances of the Si surrounded with 3Si and 1Al (3Si, 1Al) units. The shoulder peak at −117.1 ppm results from crystallographic inequivalent sites in the zeolites [43]. Based on Loewenstein’s rule, the framework the Si/Al ratio can be determined from the peak areas in the ²⁹Si MAS NMR spectrum [44]. The calculated framework Si/Al ratio of the ZM-1 and ZM-3 catalysts is 12.8 and 20.4, respectively (Table 4). The higher Si/Al ratio of the ZM-3 catalyst is in line with its higher content of EFAl compared with the ZM-1 catalyst. Indeed, the difference between the framework Si/Al ratio and the bulk Si/Al ratio (Table 2) results from the existence of EFAl in the samples.

¹H MAS NMR is a powerful technique to characterize the local environment of protons in ZSM-5, including type and strength of the acid sites as well as their location in the framework [43]. Figure 8 shows ¹H MAS NMR results of the catalysts. The quantitative deconvolution of the spectra leads to five different peaks. The peaks with a maximum at 1.7 and 3.9 ppm are related to the silanol groups (nonacidic hydroxyl) and bridging hydroxyl groups (Si(OH)Al), respectively. The peak at ca. 2.4 ppm is assigned to the hydroxyl group interacting with EFAl [43]. The peak with a maximum at 5.9 ppm is attributed to a second Brønsted acid site (see below) of the HZSM-5 catalyst [45]. The results show that the ZM-1 catalyst includes a higher concentration of the silanol groups than the ZM-3 catalyst. It is reported that the Brønsted acid sites in ZSM-5 can be categorized into two groups [46]: (i) unperturbed Brønsted acid sites that have no additional interaction with the framework and (ii) perturbed Brønsted acid sites that form hydrogen bonds with another framework oxygen atom. The hydrogen-bonding interaction helps anchor or stabilize the proton, which may reduce unwanted mobility that might be detrimental under certain reaction conditions. Moreover, the acid strength of perturbed Brønsted acid sites can be modulated because the proton is partially delocalized by the hydrogen bond [46], leading to moderate strength (i.e., not too weak, not too strong), which favors many catalytic reactions. It is well documented that strong acid sites of the ZSM-5 catalyst are the main sites responsible for coke formation and catalyst deactivation in various reactions [4,47]. The high intensity of the peak at 5.9 ppm for the ZM-1 catalyst indicates a high amount of the second Brønsted acid sites, which have a hydrogen bond with another framework oxygen atom [48]. Given the geometric flexibility and denser framework oxygen environment at the channel intersections of the MFI structure, perturbed Brønsted acid sites—where the acidic proton forms an intraframework hydrogen bond—are expected to be more probable in these regions. This hypothesis aligns with DFT models and ¹H NMR observations [46], which indicate that such hydrogen-bonded Brønsted acid sites are energetically favored in distorted or spacious environments characteristic of channel intersections. These results support the ²⁹Al MAS NMR results, indicating more Al siting in channel intersections of the ZM-1 catalyst.

The acidity properties of the samples were studied by NH₃−TPD technique (Figure 9a). The results show that the ZM-2 and ZM-3 catalysts include higher concentration and strength of the acid sites compared with the ZM-1 catalyst. The type of acid site pyridine was characterized by adsorption and desorption using in situ FTIR spectroscopy (Figure 9b). It is worth noting that preheating the catalysts (at 300 °C) reduces the impact of physisorbed water and pyridine compounds. The band at 1447 and 1625 cm⁻¹ is assigned to Lewis acid sites. The band at 1547 and 1640 cm⁻¹ is attributed to Brønsted acid sites. The band at 1490 cm⁻¹ is related to pyridine interacting with both Lewis and Brønsted acid sites [48]. The results show that the ZM-1 includes more Brønsted acid sites compared with the ZM-2 and ZM-3 catalysts (

Table 5. Acidity data of the H-ZSM-5 catalysts.

Catalyst	Acidity, μmol NH3 g−1			Peak Temperature, °C		Py-FTIR
	Weak	Strong	Total	TP1	TP2	Brønsted	Lewis	Brønsted/Lewis
ZM-1	7.2	12.5	19.7	141	341	0.34	0.17	2.03
ZM-2	16.0	20.8	36.8	152	374	0.22	0.54	0.41
ZM-3	0.6	25.0	25.6	158	357	0.11	0.12	0.91
ZM-8	0.20	3.72	3.92	173	322	0.52	2.23	0.23

), which can play a key role in catalytic activity in CO₂ hydrogenation.

3.2. Effect of H₂O/Si Ratio

The XRD results (Figure 1) show that the ZM-8 catalyst with the H₂O/SiO₂ ratio of 75 displays the typical XRD pattern of the MFI structure. The lower water content (H₂O/SiO₂ = 75) does not influence the relative crystallinity of the ZM-8 sample (Table 2). However, a further decrease in the H₂O/SiO₂ ratio (10) in the ZM-9 sample leads to an amorphous phase. The SEM images of the ZM-8 catalyst represent a coffin-shape surface morphology with a larger particle size compared with the ZM-1 catalyst (Figure 3), including the a- and b-axis length of 18.97 and 3.29 µm, respectively. Yu et al. [49] found that a lower H₂O/SiO₂ ratio affected the crystallization rate, involving a long nucleation period and fast crystal growth. Interestingly, the particles with a long a-axis show the straight channels of the ZM-8 catalyst reaching 90.5% (Table 2), which may be expected to significantly affect the shape selectivity of p-xylene in the production of the aromatic compounds in CO₂ hydrogenation. The ZM-8 catalyst represents the N₂ adsorption–desorption isotherm of type IV (Figure 4), indicating a microporous/mesoporous structure. The lower surface area (306 m² g⁻¹) and lower external surface area (83 m² g⁻¹) of ZM-8 compared with ZM-1 (Table 3) are in line with its larger particle size. The FTIR spectrum of the ZM-8 catalyst (Figure 5) shows all the characteristic bands for the MFI structure, as discussed above, which supports the XRD results. The lower water content in the synthesis gel significantly increases Lewis and Brønsted acid sites in the ZM-8 catalyst (Table 5), favoring catalytic activity in CO₂ hydrogenation through cyclization, dehydrogenation, and oligomerization reactions [50].

3.3. Catalytic Tests

The ZM-1 and ZM-8 catalysts were selected to check the catalytic activity experiments owing to their appropriate textural and acidity properties. The experiments were studied in a dual-bed fixed-bed reactor, comprising separate beds of the Na/Fe₂O₃ metal catalyst, and of the zeolite, under reaction conditions of 320 °C, 2.5 MPa, and 4000 mL g_cat⁻¹h⁻¹ (Figure 10). The iron-based catalyst promoted with 4 wt% sodium exhibited a stable and high CO₂ conversion (ca. 45%) for a long time on stream (17 h). Our previous work on the effect of alkali promoters on iron carburization for C-C coupling indicated that sodium yielded a superior activity compared with other alkali such as potassium and cesium in iron-based catalyst systems [51]. For that reason, we employed this metal catalyst to prepare tandem catalysts with zeolite. The ZM-1 and ZM-8 catalysts show low CO and CH₄ selectivity (<20%), indicating a limited reverse water–gas shift reaction for these catalysts. Xu et al. [52] reported that a hollow b-oriented ZSM-5 catalyst improved aromatic production through methane co-aromatization with propane. The literature suggests that b-oriented ZSM-5 favors the aromatization in CO₂ hydrogenation due to the improved diffusion of the aromatic compounds [53,54]. In this regard, the ZM-8 catalyst provides a higher selectivity of aromatic products (77%) compared with the ZM-1 catalyst (26%). The aromatic STY improves significantly over the ZM-8 catalyst (4 g kg_cat⁻¹h⁻¹) compared with Z-1 (2.7 g kg_cat⁻¹h⁻¹). It is accepted that Brønsted acid sites are the main active sites for aromatic production through CO₂ hydrogenation [55] as more Brønsted acidity of the ZM-8 catalyst promotes aromatic formation. Furthermore, the results can be related to the longer sinusoidal channels of the ZM-8 catalyst, providing sufficient residence time for the conversion of non-aromatic compounds to aromatics through cyclization and dehydrogenation reactions. Liu et al. [56] found that the conversion of bulky compounds over ZSM-5 catalyst was directly attributed to the role of b-oriented ZSM-5 particles because almost all mesopores in the b-axis were opened to the surface of ZSM-5 particles, so they were more accessible to bulky compounds. Most of the disordered mesopores in the bulk of the ZSM-5 particles were, however, not accessible to the bulky compounds. These results are in line with the higher BTX selectivity (22%) and lower C₉₊ selectivity (75%) over the ZM-8 catalyst (Figure 10c) compared with the ZM-1 catalyst (6% and 92%, respectively). Lim et al. [57] reported that larger ZSM-5 particles (>10 µm) favored production of BTX aromatics through the hydrodealkylation of C₉₊ aromatics. Furthermore, the high Lewis acidity of the ZM-8 catalyst can accelerate dehydrogenation reactions to convert non-aromatics to aromatics [58].

The ZM-8 catalyst shows higher p-xylene selectivity (69%) in BTX products compared with the ZM-1 catalyst (62%). Moreover, p-xylene STY increases almost 6.5 times over the ZM-8 catalyst (0.53 g kg_cat⁻¹h⁻¹), highlighting the key role of its higher fraction of straight channels (90.5%) compared with the ZM-1 catalyst. Murciano et al. [59] concluded that BTX production directly increased with the concentration of the Brønsted acid sites in CO₂ hydrogenation over the K-Fe/Al₂O₃ + HZSM-5 catalyst. Zhang et al. [54] concluded that the b-axis-oriented ZSM-5 catalyst improved p-xylene selectivity through the alkylation of benzene with methanol owing to fast diffusion in the straight channels. Wang et al. [13] reported that xylene molecules (meta and ortho) have to diffuse into the sinusoidal channels of the b-oriented ZSM-5 catalyst, favoring xylene isomerization and higher p-xylene formation. Chen et al. [60] applied computational studies based on ab initio molecular dynamic simulation and reported that the diffusion of m-xylene through the sinusoidal channels of the ZSM-5 catalyst is more difficult than its conversion to p-xylene by isomerization. Hence, the long sinusoidal channels increase the probability of the xylene isomerization reactions. p-xylene diffuses faster than m- and o-xylene in the straight channels owing to the smaller molecular size that promotes the fast exit of p-xylene as well as its selectivity [55]. Consequently, the ZM-8 catalyst provides a high percentage of p-xylene (82%) in xylene compounds over a long time on stream (17 h), which can be assigned to the excellent textural/acidity properties of the developed catalyst.

4. Conclusions

A direct hydrothermal technique was applied for the one-pot synthesis of b-oriented H-ZSM-5 with a high Al content. The results showed that TPA⁺ and NH₄⁺ cations were incorporated in the ZSM-5 framework as charge compensating agents. The NaAlO₂ as the Al source and NH₄NO₃ as the metal-free mineralizer resulted in large particles, including high surface area, high concentration of straight channels, and high concentration of acid sites located at channel intersection zone. Furthermore, the concentrated synthesis gel (low H₂O/SiO₂ ratio) favored large particle size. Therefore, the catalyst showed improved catalytic performance for long-term time on stream (17 h) in the CO₂ hydrogenation reaction, including high aromatic STY and high aromatic and p-xylene selectivity. The results represent the capability of the developed direct synthesis for the preparation of highly Al-rich ZSM-5 catalysts with high catalytic performance in CO₂ hydrogenation.