Different Routes for the Hierarchization of *BEA Zeolite, Followed by Impregnation with Niobium and Application in Ethanol and 1-Propanol Dehydration

Abstract

This study examined the hierarchical structuring of *BEA zeolite using two distinct approaches: double aluminum removal with solid ammonium hexafluorosilicate (2x-AHFS) and a solution of 0.2 M sodium hydroxide followed by 0.5 M hydrochloric acid (T-NaOH). Additionally, niobium pentoxide (Nb₂O₅) was impregnated at different loadings (5, 10, 15, and 20 wt.%) onto the hierarchized materials. Both treatments increased the SiO₂/Al₂O₃ ratio and produced crystals with domains of about the same size. The hierarchization methods generated secondary mesopores and reduced the micropores in the treated HB zeolite. The solid-state NMR analysis by ²⁷Al and ²⁹Si indicated that the 2x-AHFS treatment increased the hydrophobic character of the zeolite, while the treatment with NaOH/HCl resulted in a less hydrophobic material. A balanced quantity of Brønsted and Lewis sites was observed for all treated zeolites. Thus, these combined physicochemical characteristics of the new catalysts may explain their superior performance in the dehydration reactions. In the case of ethanol dehydration at 230 °C, the 20 wt.% Nb₂O₅ supported on the T-NaOH catalyst produced an 84% conversion and 86% selectivity for ethylene (EE), with 14% diethyl ether (DEE) as the only products. Conversely, in the 1-propanol dehydration reaction, the 20 wt.% Nb₂O₅ supported on 2x-AHFS achieved 99% conversion, producing 99% propene.

1. Introduction

Catalysis involved in the conversion of renewable resources to green chemicals has attracted significant attention over the past decade [1]. Among heterogeneous catalysts, zeolites account for more than 40% of solid catalysts used in the chemical industry due to their superior properties, such as modulated acidity, high surface area, thermal and chemical stability, and the ability to confine active metal species within their pores [2]. Monte Carlo simulation studies have shown that the structure of zeolites is inherently altered during adsorption processes, and the conversion (yield) largely depends on both the crystalline structure (e.g., 10- and 12-ring three-dimensional channel systems) and framework compositions (e.g., Si/Al ratios and the associated acidity and hydrophobicity/hydrophilicity) [3].

Zeolites are crystalline materials with pores and cavities that exhibit excellent thermal and chemical stability, Brønsted/Lewis acidity, and high catalytic activity [4]. Among the five zeolites with the greatest applicability (FER, MOR, MFI, FAU, and *BEA), *BEA zeolite stands out due to its wide micropores, intersecting channels, and 12-membered rings [5]. Despite its large pores, some reagents have difficulty accessing the active catalytic sites, and chemical reactions are limited by diffusion issues [6]. One way to overcome this problem is through the hierarchization of the zeolite, where the material acquires different types of pores. This procedure can be performed either before or after zeolite synthesis [7,8,9]. Several studies in the literature show that catalytic reactions exhibit higher conversion and selectivity for the products of interest after the hierarchization of the zeolite, including *BEA [10,11,12].

Hierarchical zeolites have interconnected pores of different sizes (bimodal micro-mesoporous structure), which reduce the steric limitations of bulky reagents and their access to catalytic sites, increase the intracrystalline diffusion rate, minimize catalyst deactivation caused by coke formation, maximize catalyst utilization, and modulate selectivity towards target products [13]. Among post-synthesis treatments of zeolites, solid-state dealumination has shown promise, as Al is removed from the structure with the reinsertion of Si into the vacancy left by Al. As a result, the Si/Al ratio increases, thereby enhancing the resistance of the zeolitic structure to extreme thermal and hydrothermal treatments [14]. On the other hand, basic leaching is a technique that removes Si from the framework (desilication) through aqueous alkaline treatment (NaOH). In addition to being promising, this technique is particularly suitable for zeolites with a high Si/Al ratio. Through the hydrolysis of Si-O-Si and Si-O-Al bonds, the created mesopores offer higher accessibility to the zeolite’s outer surface, as they are primarily generated on the outer surface or in defective areas within the structure [15,16]. According to Groen et al. [17], the framework Si/Al ratio plays a crucial role, as lattice aluminum species control the silicon extraction from the zeolite framework due to the suppressed extraction of neighboring Si species.

One of the applications of hierarchical zeolites is the conversion of biomass-derived feedstock, such as bioethanol, into renewable ethylene via a dehydration reaction operated at lower temperatures (200–500 °C). Strong solid acid sites facilitate ethylene formation, while weak acid sites promote the formation of diethyl ether [17,18]. Additionally, the use of renewable sources like propanol is important for industrial processes, such as propene production via dehydration, since propene is a building block to produce a wide range of chemicals [19]. In a previous study by Valadares et al. [14], it was demonstrated that the catalytic properties of *BEA zeolite treated with ammonium hexafluorosilicate (1×-70 mol%) were effectively tuned by incorporating niobium species, which favored the production of diethyl ether (DEE) from ethanol dehydration and furfural from xylose dehydration. The addition of niobium enhanced the selectivity for these products, providing the possibility to control them through different hierarchical processes.

Considering the above, the objective of the present study is to produce hierarchical *BEA zeolite using two different methodologies: a double dealumination with ammonium hexafluorosilicate (2×-AHFS) in the solid state and treatment in solution with sodium hydroxide, followed by treatment with hydrochloric acid (NaOH/HCl). To evaluate the modifications of the materials, they were characterized structurally and texturally using several techniques: XRD, EDXRF, FT-IR, ²⁷Al and ²⁹Si MAS NMR, SEM, nitrogen sorption at low temperature, and gaseous adsorption of pyridine followed by FT-IR analysis. All modified catalysts were tested in ethanol and 1-propanol dehydration reactions to assess the effect of desilication/dealumination on the performance of the *BEA materials.

2. Results and Discussion

2.1. Powder XRD and XRF Characterizations

The different *BEA polymorphs present similar peaks in the powder X-ray diffractograms but with different positions, which are characteristic, particularly in the low-angle region, widely accepted as the *BEA fingerprint region (2θ = 5–10°). Both sharp and broad diffraction peaks suggest structural disorder (stacking faults) of the framework due to an intergrowth of distinct but closely related polymorphs (A and B). In all samples, the main diffraction peaks were identified at 2θ = 7.80° (broad) and 22.50° (sharp), which correspond to the reflections (101) and (116) of the *BEA zeolite, respectively (Figure 1) [20,21]. The provided *BEA zeolite showed characteristics of both polymorphs in nearly equal proportion (44% A-chiral and 56% B-achiral) [22]. The modified materials showed no significant deviation in angles compared to the starting protonic zeolite (HB), indicating the preservation of the crystalline structure and no phase change. Even a double amount of ammonium hexafluorsilicate (2×-AHFS) did not compromise the zeolite network. Relative crystallinity calculations were performed using the total areas of the characteristic diffraction peaks, with the parent zeolite designated as 100% crystalline (

Table 1. Physicochemical properties of the catalysts.

Catalyst	$\frac{{\mathbf{S}\mathbf{i}\mathbf{O}}_2}{{\mathbf{A}\mathbf{l}}_2{\mathbf{O}}_3}$	%C a	2θ (°) b	β (rad) b	D (nm) c	L (nm) d	Strain e
HB	25	100	22.52	0.0105	13.1	4.6	0.0007
2x-AHFS	44	102	22.63	0.0115	11.9	5.2	0.0009
5Nb-AHFS	35	100	22.60	0.0119	11.6	6.5	0.0009
10Nb-AHFS	33	100	22.50	0.0119	11.6	6.7	0.0009
15Nb-AHFS	35	95	22.50	0.0108	12.7	6.6	0.0008
20Nb-AHFS	36	96	22.50	0.0112	12.3	6.6	0.0008
T-NaOH	39	88	22.57	0.0136	10.1	5.0	0.0012
5Nb-NaOH	42	87	22.77	0.0199	6.9	7.7	0.0023
10Nb-NaOH	41	89	22.69	0.0182	7.6	8.6	0.0020
15Nb-NaOH	41	100	22.67	0.0140	9.8	6.2	0.0013
20Nb-NaOH	38	106	22.72	0.0157	8.7	5.8	0.0015

^a Crystallinity (%) calculated by Equation (1). ^b 2θ angle and the respective full width at half maximum (FWHM). ^c Average size of crystal domain, calculated by Scherrer’s equation: D = 0.9 λ/β cosθ. ^d Average particle diameter (L), calculated by Equation (2). ^e Microstrain (ε) calculated by ε = β sin(θ)/4.

). An increase of 2% in crystallinity was observed for HB treated with 2×-AHFS, and a decrease of 12% was noted when treated with NaOH/HCl (T-NaOH). For the HB treated with NaOH, it was observed that when an extra metal source (e.g., niobium) is added above 10 wt.%, this metal will act as a healing atom, initially occupying the silanol nests to prevent further collapse of the zeolite [9,23]. The peak at 2θ = 27.1° (crystal plane 008), characteristic of the *BEA structure, decreased after NaOH treatment but not after 2x-AHFS treatment [24]. The diffractograms for the samples after the impregnation of Nb₂O₅ (Figure 1) for both treatments indicated a relative decrease in peak intensities for both peaks at 2θ = 22.5° (plane 116) and 7.8° (plane 101), since the supported Nb₂O₅ was amorphous when treated at 550 °C.

The average crystallite domain calculated by XRD was 12 nm for the 2×-AHFS and 9 nm for T-NaOH, whereas the average particle diameter obtained using BET was 7 nm and 6 nm for 2×-AHFS and T-NaOH, respectively (Table 1). In general, the values did not significantly differ in magnitude for both methods and may reflect differences in calculations [25]. However, it is interesting to note that the dealumination using NaOH resulted in smaller particles, regardless of the calculation method. In addition, The T-NaOH caused higher microstrains, from 0.0022 to 0.0029, while the AHFS treatment produced microstrains from 0.0017 to 0.0019. The treatment with NaOH probably resulted in greater intergrowth and stacking faults, potentially impacting the physicochemical properties, particularly diffusivity, due to modifications in pore connectivity.

All materials showed an increase in the SiO₂/Al₂O₃ ratio, as both treatments can remove framework Al (Table 1). A larger SiO₂/Al₂O₃ ratio corresponds to a smaller unit cell size and, thus, larger XRD angles, from 2θ = 22.52° for HB (SiO₂/Al₂O₃ = 25) to 22.57° for T-NaOH (SiO₂/Al₂O₃ = 39) and 22.63° for 2×-AHFS (SiO₂/Al₂O₃ = 44). Acid leaching can preferentially remove Al from the network and create mesopores within the zeolitic crystals. Defective sites are formed through the hydrolysis of Si-O-Al bonds, forming Si-O- defects and giving rise to extra-lattice Al species (EFAl) [24].

The Nb₂O₅ loading was obtained based on Nb analysis determined by energy dispersive X-ray fluorescence (EDXRF), which used the QualiQuant method of fundamental standards (Version 1.2, 2006, Shimadzu, Kyoto, Japan). The theoretical values were close to the real ones (deviations of 2–3%). Therefore, the nominal values were used here for simplicity.

2.2. FT-IR Spectroscopy

In relation to the FT-IR spectra, all catalysts showed characteristic bands at approximately 1220 cm⁻¹, which corresponds to the external symmetric vibration of the SiO₄ tetrahedron, and at approximately 1089 cm⁻¹ (Figure 2), which corresponds to the asymmetric vibration between the tetrahedral atom (T, Si or Al) and the oxygen atoms (O-T-O bonds) [26]. After the modifications, these bands shifted to higher wavenumbers, which is an expected phenomenon in hierarchization due to the increase in the SiO₂/Al₂O₃ ratio [27]. The band around 946 cm⁻¹ was also identified, corresponding to the vibration of the Si-O- bond [14,27]. This band was more prominent in materials treated with NaOH than in those treated with 2x-AHFS. Additionally, the band at 946 cm⁻¹ is sensitive to the incorporation of niobium into the framework. When Nb is incorporated into the zeolite structure, the wavenumber shifts to higher values (in the range of 960–970 cm⁻¹), according to the literature [14,28]. This shift was not observed, reinforcing the presence of only niobium pentoxide species. Finally, the absorption bands observed around 628, 572, and 526 cm⁻¹ correspond to the vibrations of the 4-, 5-, and 6-membered rings of the *BEA zeolite, respecitvely, indicating that the modifications did not significantly compromise the zeolite framework [14,28,29]. Additionally, a large band from 3550 to 3750 cm⁻¹ was observed for the samples prepared under an inert atmosphere with Nujol (Figure S1). A band at 3610 cm⁻¹ is assigned to bridging hydroxyl groups [Si-O(H)-Al], and another at approximately 3650 cm⁻¹ is related to the presence of extra-framework Al-OH, as reported in the literature [9,30].

2.3. Textural Properties Using N₂ Adsorption/Desorption Isotherms at −196 °C

Figure S2 shows the N₂ physisorption isotherms (adsorption/desorption). All samples showed a combination of type I(a) and IV(a) isotherms with a hysteresis loop at higher relative pressures (p/p₀ > 0.7) that indicated the presence of a secondary mesoporous structure. The specific surface area and pore volume data are presented in

Table 2. Textural parameters for the catalysts.

Catalyst	SBET a (m2/g)	SExt b (m2/g)	SMicro c (m2/g)	SMeso d (m2/g)	VMicro e (cm3/g)	Vp f (cm3/g)	VMeso g (cm3/g)
HB	649	190	459	188	0.19	0.85	0.66
2x-AHFS	577	193	384	211	0.16	0.87	0.71
5Nb-AHFS	462	170	292	193	0.12	0.67	0.55
10Nb-AHFS	448	156	292	179	0.12	0.63	0.51
15Nb-AHFS	484	163	321	199	0.13	0.67	0.54
20Nb-AHFS	512	183	330	211	0.13	0.77	0.64
T-NaOH	607	220	387	239	0.16	0.97	0.81
5Nb-NaOH	391	222	169	151	0.07	0.68	0.61
10Nb-NaOH	347	138	209	137	0.08	0.51	0.43
15Nb-NaOH	484	163	320	198	0.13	0.76	0.63
20Nb-NaOH	488	168	321	199	0.13	0.76	0.63

^a Specific surface area obtained by BET method. The standard error (2σ) was ±2 m² g⁻¹. ^b External surface obtained by t-plot method. ^c Microporous area obtained by t-plot method. ^d Mesoporous area obtained by BJH method. ^e Microporous volume obtained by t-plot method. ^f Total pore volume obtained by the amount of gas adsorbed at p/p₀ = 0.98. ^g Mesoporous volume (Vp − VMicro).

. The specific surface area (S_BET), mesopore area (S_Meso), total pore volume (V_p), and mesopore volume (V_Meso) increased after both treatments in the following order: T-NaOH > 2x-AHFS > HB. This indicates an effective process of hierarchization of the zeolite BEA using both treatments. However, the addition of niobium forming Nb₂O₅ did not produce a crescent increase in all those values. For the Nb-NaOH catalysts, low loading of Nb₂O₅ (5 to 10 wt.%) initially decreased the BET area and caused a decrease in V_p and hence in the micropore and mesopore volumes. This can be attributed to the inclusion of small niobium clusters inside the zeolite pores, which precluded the complete filling of the original pores of T-NaOH zeolite. It is known that low niobium loading forms preferentially smaller clusters or isolated species on the surface of different supports [31,32]. Conversely, when the loading was increased to 15 to 20 wt.%, there was a subsequent increase in both the BET surface area and the total surface area of the catalysts. The development of larger clusters of Nb₂O₅ on the external surface creates an overlayer that effectively contributes to the total area of the material. This has been observed for Nb₂O₅ supported on silica-alumina catalysts [33]. The trend on the textural data observed for loading Nb on 2x-AHFS was generally in parallel with that observed on T-NaOH, which is consistent with the obtained isotherms. Some of those cited trends were plotted as a function of the treatment for easier comparison and are presented in Figure S3.

Thus, it can be inferred that the impregnation of HB zeolite after these hierarchization processes primarily generates extra-framework niobium oxide species on its surface, with the overlayer thickness depending on the niobium loading, as pointed out in the literature [34]. Moreover, the textural data showed that the diffusional constraints were alleviated by coupling the intrinsic microporosity with an auxiliary mesopore framework of inter- and intracrystalline structures. The model of silicon and aluminum dissolution is the main mechanism behind the chemical Si environments and mesopores formation [35,36,37,38].

2.4. SEM Images and Analysis

The SEM micrographs of the treated zeolites and their niobium pentoxide (Nb₂O₅) impregnation are shown (Figure 3, and with EDX analyses in Figure S4) and compared to those of the HB zeolite. The protonic *BEA zeolite exhibited a distinct morphology of nanosized aggregates, as previously described in the literature [9]. After treatment with 2x-AHFS, the formation of a nearly spherical ball morphology with a uniform size was observed. The treatment using NaOH/HCl resulted in a material with morphological characteristics similar to those of HB. These images are consistent with the XRD patterns, which showed no significant differences among the materials. The impregnation of Nb₂O₅ onto both treated zeolites resulted in a rougher particle surface, but with characteristics analogous to those of the parent HB zeolite.

2.5. ²⁷Al and ²⁹Si MAS NMR Spectroscopy

In the ²⁷Al MAS NMR spectra, all catalysts showed signals around −22 to 22 ppm for octahedral Al and 40 to 80 ppm for tetrahedral Al in the framework (Figure 4). The intense signal at 1.4 ppm present in the T-NaOH zeolite has been assigned to octahedral species that have fewer adsorbed water molecules, while the broadening of the signals in this region indicates a greater degree of hydration of the catalysts [39,40]. According to Groen et al. [17], the formation of mesopores in *BEA requires the dissolution of both Si and Al, with part of the removed Al being re-incorporated into the solid. This re-incorporation may be facilitated in the presence of niobium. For the treatment with 2x-AHFS, the ²⁷Al MAS NMR spectra showed a similar distribution of tetrahedral and octahedral Al as the HB zeolite. Niobium impregnation led to an increase in the relative area of tetrahedral Al for both treatments, in the range of 5 to 15% Nb₂O₅ (

Table 3. Relative distribution of tetrahedral (Al-T_d) and octahedral (Al-O_h) aluminum, according to the ²⁷Al MAS NMR spectra.

Catalyst	δ Al-Td (ppm)	Area (%)	δ Al-Oh (ppm)	Area (%)
HB	56	62	−4.0	38
2x-AHFS	56	65	−4.0	35
5Nb-AHFS	57	70	−0.9	30
10Nb-AHFS	56	70	−6.8	30
15Nb-AHFS	57	69	−5.2	31
20Nb-AHFS	56	81	−4.3	19
T-NaOH	57	76	1.2	24
5Nb-NaOH	57	89	−1.4	11
10Nb-NaOH	57	91	−2.2	9
15Nb-NaOH	57	91	−7.2	9
20Nb-NaOH	57	79	−4.9	21

). For samples with 20% Nb₂O₅, the amount of tetrahedral Al reached a plateau around 80%. These behaviors suggest that niobium plays an important role in the reorganization of the zeolitic structure after hierarchical treatments.

In the ²⁹Si MAS NMR spectra (Figure 5), the HB zeolite showed a signal at −102 ppm, which is related to the Q³ environment (Si(1Al)). This signal almost disappears in the sample treated with 2x-AHFS. The signals at −111 and −115 ppm correspond to the Q⁴ environments (Si(0Al)), with the latter (−115 ppm) not present in the HB sample but appearing markedly in the 2x-AHFS treated sample and in those impregnated with niobium. This suggests that Al removal from the zeolite promotes the reorganization of Si in the sample, leading to the formation of more Q⁴-type environments [14,30,39]. The signal at −115 ppm can be attributed to the formation of silica (nest silanol) [14,30,35], which has a chemical shift in the same region and is more prominent in the treatment with 2x-AHFS. On the other hand, for the T-NaOH sample, the Q³ environment remained practically unchanged relative to the HB sample. In addition, the Q⁴ environment at −111 ppm remained constant, and the signal at −115 ppm did not change even with the addition of niobium. Therefore, in the treatment with 2x-AHFS, nest silanol is incorporated into the framework, which increases the Q⁴ environment at −115 ppm, corroborating the increase in crystallinity observed by XRD and the higher SiO₂/Al₂O₃.

Table 4. Relative distribution of Q⁴ (Si(0Al)) and Q³ (Si(1Al)) silicon environments based on deconvolution of ²⁹Si MAS NMR spectra.

Catalyst	Q3 (ppm)	Area (%)	Q4 (ppm)	Area (%)	Q4 (ppm)	Area (%)
HB	−103	20	−111	67	−115	13
2x-AHFS	−102	8	−112	83	−115	9
5Nb-AHFS	−103	15	−112	74	−115	11
10Nb-AHFS	−103	20	−112	72	−115	8
15Nb-AHFS	−103	13	−112	78	−115	9
20Nb-AHFS	−103	14	−112	76	−114	10
T-NaOH	−102	26	−111	67	−114	7
5Nb-NaOH	−103	23	−112	66	−115	12
10Nb-NaOH	−104	29	−110	50	−113	21
15Nb-NaOH	−103	10	−112	84	−115	6
20Nb-NaOH	−103	18	−112	73	−115	10

and Figure S5 provide a detailed presentation of the distribution environments derived from the deconvolution of the spectra.

2.6. Acidity of the Catalysts

Figure 6 presents the FT-IR spectra of pyridine adsorbed on the catalysts. The HB zeolite showed stronger intensity in the three bands between 1580 and 1420 cm⁻¹. These bands correspond to the interaction between pyridine and Brønsted sites (1540 cm⁻¹), the interaction with Lewis sites (1445 cm⁻¹), and a band at 1490 cm⁻¹, which is assigned to the interaction with both types of acid sites [41,42]. Qualitative analysis revealed that spectra from 2x-AHFS and T-NaOH treatments had the same bands, but with lower intensity compared to HB. After Nb₂O₅ impregnation, the 5% and 10% loadings showed a minor decrease in intensity, whereas the 15% and 20% loadings exhibited a relative increase. This may be related to the quantity of acid sites in these catalysts. The addition of niobium to the treated zeolites initially decreases the total amount of acid sites due to partial blocking access to the original acid sites of HB, which are in the micropores [14]. However, as the niobium loading increases, the formed Nb₂O₅ overlayer increases the total accessible sites (including the niobium contribution) on the new surface. The key point is that these acid sites facilitate the catalytic activity of these materials in dehydration reactions.

Based on the FT-IR spectra, it was also possible to calculate the ratio between the areas corresponding to the Brønsted and Lewis sites (

Table 5. The relative Brønsted to Lewis ratio (B/L) obtained by FT-IR using Py adsorbed on the catalysts.

Catalyst	B/L
HB	1.1
2x-AHFS	1.2
5Nb-AHFS	1.1
10Nb-AHFS	1.1
15Nb-AHFS	1.1
20Nb-AHFS	1.1
T-NaOH	1.1
5Nb-NaOH	1.2
10Nb-NaOH	1.2
15Nb-NaOH	1.1
20Nb-NaOH	1.2

). As a result, the 2x-AHFS treatment showed a higher ratio value than T-NaOH. It is possible that the AHFS treatment removed only the extra-lattice Al and introduced Si atoms into the material. The T-NaOH treatment broke the Si-O bonds in the framework, followed by the removal of Al using an HCl solution, which led to a decrease in the Brønsted sites of the zeolite. Impregnation with Nb₂O₅ kept the ratio almost constant. This behavior provides evidence that Nb₂O₅ can influence the restructuring of Brønsted and Lewis sites. Niobium, with electronegativity similar to that of Al, may promote the interaction of pyridine with Brønsted sites and act as a coordinative center for pyridine.

2.7. Ethanol Catalytic Dehydration

As pointed out in the literature, olefin production from (bio)alcohol dehydration is an important industrial process for the potential use of raw materials derived from biomass, substituting petroleum-based materials [43,44]. For the complete transformation (conversion) of ethanol, there are four types of reactions: (i) dehydration, (ii) dehydrogenation, (iii) dehydrogenation coupling, and (iv) hydrogenolysis [45,46,47,48]. The final selectivity can be controlled by a catalyst with different active sites (e.g., solid acid, bifunctional). When using a solid acid, the number, strength, and distribution of Brønsted and Lewis sites are critical, as is the accessibility to these sites.

As previously described, the dehydration reaction of ethanol was performed at 230 and 250 °C. The results are described from two perspectives: alcohol conversion and selectivity. Figure 7 shows the results at 230 °C. Compared to the parent zeolite (HB), which showed 91% conversion, both treated zeolites showed lower conversion (2x-AHFS = 51% and T-NaOH = 64%). Modifying the surface of treated zeolites with Nb₂O₅, it was observed a distinct behavior. The 2x-AHFS had an average conversion of around 47%, whereas the T-NaOH samples generally increased the conversion with the niobium loading, reaching a maximum of 84% with 20Nb-NaOH. Looking at the selectivity (Figure 7), we noted that both treated zeolites loaded with Nb₂O₅ showed only ethylene (EE) and diethyl ether (DEE) as the products. Although the HB zeolite showed the highest conversion and selectivity for EE and DEE, there is a minor formation (about 1%) of other products. Probably, because of the higher strength of Brønsted sites in HB, there is also an increase in parallel reactions compared to the modified zeolites. The most active catalysts were 20Nb-AHFS which showed 55% conversion, with selectivity of 56% (EE) and 41% (DEE), while 20Nb-NaOH had 84% conversion and selectivity of 86% (EE) and 14% (DEE).

The same analysis approach was employed to run the reaction at 250 °C. The results for conversion and selectivity are in Figure 8. In general, all the catalytic conversions increased with prominent results for HB, which reached 96%, but again with 3% selectivity for other products. The treated HB zeolite displayed an increased conversion (52% for 2x-AHFS and 74% for T-NaOH). The insertion of niobium leads to the maximum conversion reaching 84%, with 82% selectivity for EE and 18% for DEE using a 20Nb-NaOH catalyst. Therefore, it can be concluded that although the highest conversion was achieved with HB, the exclusive products (EE and DEE) with high selectivity towards EE are more advantageous for the treated zeolites with Nb₂O₅, particularly the 20Nb-NaOH sample. As the reaction temperature increased from 230 to 250 °C, the conversion rate of the catalyst 20Nb-NaOH tended to level up with HB zeolite with the benefits of only EE and DEE products.

Dehydration reaction results showed that factors such as catalyst morphology and the presence of mesoporosity may enhance mass transfer efficiency, enabling more effective access of alcohol to the active sites, and resulting in higher conversions. The strength and distribution of Brønsted/Lewis sites affected the selectivity for alkene versus ether. The presence of Nb₂O₅ on these catalysts clearly modulates these latter parameters. As we observed earlier [14], the addition of an overlayer of Nb₂O₅ weakens the catalyst compared to the original HB sites. However, selectivity control is clearly dependent on the dealumination treatment. HB treated with 1x-AHFS was more selective for DEE than EE, while the opposite was observed for the 2x-AHFS treatment. The higher efficiency for producing EE with 20Nb-NaOH can be attributed to a combination of increased Brønsted sites from the addition of niobium and better accessibility to these sites due to the mesoporous contribution in this catalyst.

2.8. 1-Propanol Catalytic Dehydration

Propene is one of the fundamental products to support petrochemical processes [19]. Producing propene from renewable sources like propanol can be important for an economy based on lower carbon emissions using biomass [49]. Thus, our catalysts were evaluated in this reaction. Figure 9 presents the results of conversion and selectivity for the dehydration process at 230 °C. The conversion exceeded 88% for all catalysts. For instance, HB converted 90% of 1-propanol, while the treated zeolites showed 96% for 2x-AHFS and 99% with T-NaOH. This indicates that both treatments were efficient in producing more active catalysts than the parent zeolite. The addition of Nb₂O₅, independent of the treatment, maintains the level of conversion above 96%, which demonstrates the efficiency of these catalysts over HB. Actually, the effect on selectivity is much more interesting. HB zeolite showed only 18% selectivity for propene (PP), whereas the 2x-AHFS and T-NaOH had 100 and 89%, respectively. The addition of Nb₂O₅ brought about the same level (99%) of selectivity for PP for treatment with 2x-AHFS and an average of 90% for T-NaOH. Clearly, the main enhancement in the high selectivity for PP was achieved by the dealumination treatment of the parent zeolite (HB).

One important factor that could explain the higher selectivity for propene in these catalysts may be related to the increased mesoporous area after the dealumination treatments. As we have discussed, this increased area may affect the accessibility of 1-propanol to the Brønsted acid sites of these zeolites, which can better accommodate intermediates based on dimers or trimers of the 1-propanol adsorbed on the acid sites. This is consistent with mechanistic studies in the literature on 1-propanol dehydration over HZSM-5 zeolite [50].

3. Materials and Methods

3.1. Hierarchization of *BEA Zeolite

Zeolite NH₄BEA, obtained from Zeolyst International (CP814E, mole ratio SiO₂/Al₂O₃ = 25, Conshohocken, PA, USA), was calcined (8 h at 550 °C) forming a proton sample (HB), which underwent dealumination through two different treatments:

(i)

Solid state hierarchization: a fraction of HB was subjected to solid-state dealumination using ammonium hexafluorosilicate (AHFS, 98%, Aldrich, Burlington, MA, USA) with the intention of removing a theoretical percentage of 70 mol% of Al from its crystal lattice by successive (2x) for comparison purposes. The solids were placed in an agate mortar and pestle, and mixed for 10 min, followed by placing the mixture in a desiccator containing a saturated ammonium chloride solution (>99.5%, Sigma-Aldrich, Burlington, MA, USA) at atmospheric pressure. After 24 h, the mixture was heated in a muffle furnace (3 h, 190 °C), followed by washing with ammonium acetate solution and deionized water (Milli-Q), both at room temperature. Finally, the mixture was dried in an oven (24 h, 120 °C) and calcined (8 h, 550 °C) [14].

(ii)

Hierarchization in base and acid solutions: the HB zeolite was treated with a 0.2 M sodium hydroxide solution (97%, Aldrich, Burlington, MA, USA), under magnetic stirring at 75 °C, for 4 h. Subsequently, this mixture was washed with deionized water for 1 h, also at 75 °C. Following this step, the resulting material was treated with a 0.5 M hydrochloric acid solution (37%, Aldrich, Burlington, MA, USA) under the same conditions as the base treatment. Finally, the same washing procedure was conducted, and the resulting zeolite was placed in a crucible, dried in an oven for 12 h at 120 °C, and calcined for 8 h at 550 °C.

3.2. Impregnation of Niobium

The insertion of niobium pentoxide into the materials prepared by the methodologies was achieved through aqueous impregnation. The samples were dried at 200 °C for 4 h under vacuum to remove adsorbed water molecules. Subsequently, the mass of the dry materials was measured, and based on these data, the amount of Nb₂O₅ to be used (5, 10, 15, and 20 wt.%) was determined. Niobium ammonium oxalate (99%, CBMM, Araxá, Brazil) was used as the niobium source and was completely dissolved in deionized water. The solutions containing niobium and zeolite were placed under magnetic stirring at 90 °C until the solvent had fully evaporated. Finally, the resulting materials were kept dry in an oven for 12 h at 120 °C, followed by calcination for 8 h at 550 °C.

Table 6. Nomenclature for the studied *BEA zeolite.

Code	Description
HB	Protonic *BEA zeolite
2x-AHFS	HB dealuminated twice 70 mol%
5Nb-AHFS	HB dealuminated twice 70 mol% and impregnated with 5 wt.% of Nb2O5
10Nb-AHFS	HB dealuminated twice 70 mol% and impregnated with 10 wt.% of Nb2O5
15Nb-AHFS	HB dealuminated twice 70 mol% and impregnated with 15 wt.% of Nb2O5
20Nb-AHFS	HB dealuminated twice 70 mol% and impregnated with 20 wt.% of Nb2O5
T-NaOH	HB treated with NaOH and HCl
5Nb-NaOH	HB treated with NaOH and HCl and impregnated with 5 wt.% of Nb2O5
10Nb-NaOH	HB treated with NaOH and HCl and impregnated with 10 wt.% of Nb2O5
15Nb-NaOH	HB treated with NaOH and HCl and impregnated with 15 wt.% of Nb2O5
20Nb-NaOH	HB treated with NaOH and HCl and impregnated with 20 wt.% of Nb2O5

presents the nomenclature used to identify the different samples.

3.3. Methods of Characterization

X-ray diffraction (XRD) data were obtained using a powder diffractometer (Panalytical, model Empyrean, Westborough, MA, USA) emitting radiation from a copper tube (Kα = 1.5406 Å), at 40 kV and 45 mA, with a scanning rate of 2° per minute (2θ range from 2° to 60°, with a step size of 0.02°). The crystallinity (%C) was obtained by a comparison of the XRD pattern of the standard HB, calculated by integration of the area under peaks (2θ = 5 to 60°), according to Equation (1):

$$\mathrm{\%} \mathrm{C}={\displaystyle \frac{\sum \mathrm{d}\mathrm{e}\mathrm{a}\mathrm{l}\mathrm{u}\mathrm{m}\mathrm{i}\mathrm{n}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{d} \mathrm{z}\mathrm{e}\mathrm{o}\mathrm{l}\mathrm{i}\mathrm{t}\mathrm{e} \mathrm{p}\mathrm{e}\mathrm{a}\mathrm{k}\mathrm{s}}{\sum \mathrm{H}\mathrm{B} \mathrm{p}\mathrm{e}\mathrm{a}\mathrm{k}\mathrm{s}}}\times 100.$$

(1)

Energy dispersive X-ray fluorescence (EDXRF) spectrometer (Shimadzu, model EDX 720, Kyoto, Japan) was employed to determine the quantities of silicon, aluminum, and niobium. The equipment uses rhodium (Rh) as an X-ray target and operates between 15 and 50 kV. The samples were prepared with polypropylene film and were analyzed under vacuum focused on calculating the total silica/alumina ratio.

All infrared spectra were acquired in a FT-IR spectrometer (Thermo Scientific spectrometer, Nicolet, model 6700, Waltham, MA, USA), with 512 scans and 4 cm⁻¹ spectral resolution. They were analyzed in the range of 4000 to 430 cm⁻¹, generally using 0.6 wt.% of the catalyst diluted in dried KBr pellets (>99%, Merck, Rahway, NJ, USA).

The textural data were obtained through gaseous N₂ physisorption at −196 °C in a surface analyzer equipment (Micromeritics, model ASAP 2020C, Norcross, GA, USA). Before the analysis, samples were degassed with evacuation (target pressure of 50 μm Hg) at 300 °C for 4 h. The equations of BET (Brunauer, Emmet, and Teller) in the range P/P₀ = 0 to 0.1, t-Plot and BJH (Barrett, Joyner, and Halenda) were used to describe the experimental isotherms.

The average particle diameter was calculated using Equation (2).

$$\mathrm{L}(\mathrm{n}\mathrm{m})={\displaystyle \frac{6000}{\mathrm{S}({\mathrm{m}}^2/\mathrm{g}).\rho (\mathrm{g}/\mathrm{c}{\mathrm{m}}^3)}}.$$

(2)

Scanning electron microscopy (SEM) was performed using a scanning electron microscope JEOL (model JSM, Tokyo, Japan) equipped with a secondary electron detector (LED, low energy detector), under high vacuum and a voltage of 15 kV and magnifications ranging from 100 to 10,000 times.

Solid-state nuclear magnetic resonance spectra were acquired with magic angle spinning rotation (MAS NMR) in a spectrometer (Bruker, model Avance III HD Ascend, 14.1 T, 600 MHz for ¹H, Karlsruhe, Germany) using 2- or 4-mm CP/MAS probes. Each catalyst was packed inside a zirconia rotor and specific calibration parameters were adopted for each nucleus: (i) ²⁷Al MAS NMR (156.4 MHz); 10 kHz spin rate; 0.4 µs (π/20) pulse duration; 1 s interval between pulses; 2000 acquisitions; external reference: hexa(aqua)aluminum(III) trichloride salt, [Al(H₂O)₆]Cl₃ (δ = 0 ppm). (ii) ²⁹Si MAS NMR (119.3 MHz); 10 kHz spin rate; 4.25 µs (π/2) pulse duration; 20 s interval between pulses; 3072 acquisitions; external reference: tetramethylsilane, TMS, Si(CH₃)₄ (δ = 0 ppm).

The relative distribution of aluminum atoms in each chemical environment observed in the ²⁷Al spectrum (octahedral–O_h or tetrahedral–T_d) was calculated according to Equation (3), whereas the Si/Al ratio of the zeolitic framework was calculated according to Equation (4) [43,44], which are related to the intensity of the signals referring to the chemical environments Q⁴, Q³, Q², and Q¹ after the deconvolution of the ²⁹Si spectrum (using Gaussian Lorentz function, LB = 10, Python, Version: 3.9.7 (default, 16 September 2021, 16:59:28) [MSC v.1916 64 bit (AMD64)], Comments: Enhanced by IPython 8.18.1, Python Software Foundation, Wilmington, DE, USA).

$$\mathrm{\%}\mathrm{A}\mathrm{l}({\mathrm{O}}_{\mathrm{h}} \mathrm{o}\mathrm{r} {\mathrm{T}}_{\mathrm{d}})={\displaystyle \frac{\mathrm{S}\mathrm{i}\mathrm{g}\mathrm{n}\mathrm{a}\mathrm{l} \mathrm{a}\mathrm{r}\mathrm{e}\mathrm{a} \mathrm{o}\mathrm{f} \mathrm{i}\mathrm{n}\mathrm{t}\mathrm{e}\mathrm{r}\mathrm{e}\mathrm{s}\mathrm{t}}{\sum \mathrm{S}\mathrm{p}\mathrm{e}\mathrm{c}\mathrm{t}\mathrm{r}\mathrm{u}\mathrm{m} \mathrm{a}\mathrm{r}\mathrm{e}\mathrm{a}\mathrm{s}}},$$

(3)

$${\left({\displaystyle \frac{\mathrm{S}\mathrm{i}}{\mathrm{A}\mathrm{l}}}\right)}_{\mathrm{l}\mathrm{a}\mathrm{t}\mathrm{t}\mathrm{i}\mathrm{c}\mathrm{e}}={\sum }_{\mathrm{n}=0}^4{\displaystyle \frac{\mathrm{n}}{4}}{\mathrm{I}}_{\mathrm{S}\mathrm{i}\left(\mathrm{n}\mathrm{A}\mathrm{l}\right)},$$

(4)

where I_Si(nAl) corresponds to the signal intensity related to the environment of Si bonded to n Al atoms. As the signal intensity is proportional to the number of Si atoms in each chemical environment, the total number of Si is proportional to the sum of all chemical environments (numerator). According to Löwenstein’s rule, there are no Al-O-Al groups, so each aluminum is surrounded by four Si atoms, so each Si(nAl) group corresponds to ¼ of aluminum atoms (denominator).

Acidity was measured by pyridine adsorption. Before gaseous pyridine (Py) adsorption, each sample (~20 mg) was placed in an aluminum crucible and inserted into a glass tube inside a tubular furnace (Thermolyne, model F21100, Cole-Parmer, Vernon Hills, IL, USA). The catalysts were dehydrated in dried N₂ flow (100 mL min⁻¹) at 300 °C for 1 h. Then, the system was cooled in situ to 150 °C to initiate gaseous Py (>99.8%, Sigma-Aldrich, USA) passage through samples for 1 h. After that, the temperature was held at 150 °C in an N₂ (5.0, White Martins, São Paulo, Brazil) environment for 1 h to remove any physically adsorbed Py. Immediately after cooling the system, each sample was analyzed using thermal analysis and FT-IR (the sample was prepared with a mixture of 10:170 wt.% sample/KBr). Other details can be found elsewhere [14].

3.4. Catalytic Dehydration Reactions

Dehydration reactions of ethanol (99.5%, Dinamica, São Carlos, Brazil) or 1-propanol (99.5%, Sigma-Aldrich, St. Louis, MO, USA) were evaluated by three alcohol injections under 10 mg of catalyst in a pulse microreactor coupled to a gas chromatograph with a flame ionization detector (Shimadzu GC-FID, model 2010, Kyoto, Japan), equipped with a Shimadzu CBP1 PONA column (M50-042, 50 m × 0.15 m × 0.33 μm). Reactions were carried out at temperatures of 230 or 250 °C under these conditions: alcohol injection volume: 0.3 μL; pressure: 100 kPa; total flow: 10.8 mL/min; column flow: 0.1 mL/min; linear velocity: 6.5 cm/s; purge flow: 1 mL/min; split rate: 100; column temperature: 50 °C; flame temperature: 250 °C. The calculations for the conversion of ethanol or 1-propanol, and selectivity for ethylene (EE), dimethyl ether (DME), and propene (PP) were defined by Equations (5)–(8), where n is the number of moles of the reactant.

$$\mathrm{Conversion} (\%)={\displaystyle \frac{{{\mathrm{n}}_{\mathrm{e}\mathrm{t}\mathrm{h}\mathrm{a}\mathrm{n}\mathrm{o}\mathrm{l}}}_{\mathrm{i}\mathrm{n}\mathrm{i}\mathrm{t}\mathrm{i}\mathrm{a}\mathrm{l}}-{{\mathrm{n}}_{\mathrm{e}\mathrm{t}\mathrm{h}\mathrm{a}\mathrm{n}\mathrm{o}\mathrm{l}}}_{\mathrm{f}\mathrm{i}\mathrm{n}\mathrm{a}\mathrm{l}}}{{{\mathrm{n}}_{\mathrm{e}\mathrm{t}\mathrm{h}\mathrm{a}\mathrm{n}\mathrm{o}\mathrm{l}}}_{\mathrm{i}\mathrm{n}\mathrm{i}\mathrm{t}\mathrm{i}\mathrm{a}\mathrm{l}}}}\times 100.$$

(5)

$$\mathrm{Selectivity} (\%)={\displaystyle \frac{n_{EE \mathrm{or} } n_{DME}}{{{\mathrm{n}}_{\mathrm{e}\mathrm{t}\mathrm{h}\mathrm{a}\mathrm{n}\mathrm{o}\mathrm{l}}}_{\mathrm{i}\mathrm{n}\mathrm{i}\mathrm{t}\mathrm{i}\mathrm{a}\mathrm{l}}-{{\mathrm{n}}_{\mathrm{e}\mathrm{t}\mathrm{h}\mathrm{a}\mathrm{n}\mathrm{o}\mathrm{l}}}_{\mathrm{f}\mathrm{i}\mathrm{n}\mathrm{a}\mathrm{l}}}}\times 100.$$

(6)

$$\mathrm{Conversion} (\%)={\displaystyle \frac{{{\mathrm{n}}_{\mathrm{p}\mathrm{r}\mathrm{o}\mathrm{p}\mathrm{a}\mathrm{n}\mathrm{o}\mathrm{l}}}_{\mathrm{i}\mathrm{n}\mathrm{i}\mathrm{t}\mathrm{i}\mathrm{a}\mathrm{l}}-{{\mathrm{n}}_{\mathrm{p}\mathrm{r}\mathrm{o}\mathrm{p}\mathrm{a}\mathrm{n}\mathrm{o}\mathrm{l}}}_{\mathrm{f}\mathrm{i}\mathrm{n}\mathrm{a}\mathrm{l}}}{{{\mathrm{n}}_{\mathrm{p}\mathrm{r}\mathrm{o}\mathrm{p}\mathrm{a}\mathrm{n}\mathrm{o}\mathrm{l}}}_{\mathrm{i}\mathrm{n}\mathrm{i}\mathrm{t}\mathrm{i}\mathrm{a}\mathrm{l}}}}\times 100.$$

(7)

$$\mathrm{Selectivity} (\%)={\displaystyle \frac{n_{PP } }{{{\mathrm{n}}_{\mathrm{p}\mathrm{r}\mathrm{o}\mathrm{p}\mathrm{a}\mathrm{n}\mathrm{o}\mathrm{l}}}_{\mathrm{i}\mathrm{n}\mathrm{i}\mathrm{t}\mathrm{i}\mathrm{a}\mathrm{l}}-{{\mathrm{n}}_{\mathrm{p}\mathrm{r}\mathrm{o}\mathrm{p}\mathrm{a}\mathrm{n}\mathrm{o}\mathrm{l}}}_{\mathrm{f}\mathrm{i}\mathrm{n}\mathrm{a}\mathrm{l}}}}\times 100$$

(8)

4. Conclusions

This work focuses on the hierarchization of *BEA zeolite using two methods: aluminum removal with solid ammonium hexafluorosilicate via two sequential steps (2x-AHFS) and desilication with NaOH followed by hydrochloric acid (T-NaOH). In general, the second treatment (T-NaOH) was more efficient considering the results obtained in the ethanol dehydration reaction. Top–down syntheses involving base and acid treatments are more readily implemented at an industrial scale, allowing interconnectivity and accessibility from the external surface of the zeolite crystal. We presented two beneficial effects of post-synthetic hierarchization of BEA zeolite followed by impregnation of Nb₂O₅ (5, 10, 15, and 20 wt.%) for the dehydration of ethanol and 1-propanol. The hierarchization of the *BEA zeolite by 2x-AHFS and T-NaOH promoted an increase in mesopores and the SiO₂/Al₂O₃ ratio, with only a slight decrease in the average size of crystal domains. This is important since a decrease in size could result in lower selectivity of the zeolite. An increase of 2% in crystallinity was observed for 2x-AHFS, and a decrease of 12% for T-NaOH, compared to HB zeolite. Moreover, ²⁷Al and ²⁹Si MAS NMR results indicated that a more hydrophobic catalyst was formed after the treatment of 2x-AHFS due to an increase of 12% in Q⁴ Si environments and a decrease in octahedral Al. On the other hand, T-NaOH had the highest relative quantity of tetrahedral Al (76%) and the lowest octahedral Al (24%), with a 6% decrease in the Q⁴ Si environments, leading to a less hydrophobic catalyst. The gradual addition of Nb₂O₅ tended to balance the number of tetrahedral Al (e.g., 80% of Al (Td) for 20 wt.% Nb on T-NaOH). Based on these combined physicochemical properties, the performance of the catalysts in the dehydration reactions can be explained. For ethanol dehydration at 230 °C, the best catalyst was 20 wt.% Nb₂O₅ supported on T-NaOH zeolite (20Nb-NaOH), which produced 84% conversion with 86% selectivity for ethylene (EE) and 14% diethyl ether (DEE) only. This is advantageous over HB, which showed 93% conversion but produced more than 1% of different by-products. On the other hand, the 20 wt.% Nb₂O₅ supported on 2x-AHFS zeolite (20Nb-AHFS) was the most active for 1-propanol dehydration at 230 °C, with about 99% conversion and selectivity to propene. This performance surpassed that of HB zeolite, which showed 90% conversion but only 18% selectivity for propene.