Structural and Textural Properties of Al/Cu- and Al/Zn-Pillared Clays for Ethanol Conversion

Abstract

A montmorillonite sample was pillared using mixed solutions of Al/Cu and Al/Zn as the pillaring agent. Al/Cu- and Al/Zn-pillared clays were applied in an ethanol conversion reaction. The catalysts prepared from montmorillonite were characterized using X-ray diffraction (XRD), X-ray fluorescence (XRF), 27Al nuclear magnetic resonance spectroscopy (MAS-NMR), and textural analysis by adsorption of N₂. The synthesized materials showed basal spacing values around 1.6 nm for the Cu samples and 1.8 nm for the Zn samples. With the pillarization, textural properties such as specific area and micropore volume were optimized, and all samples showed an increase in micropore volume as well as a narrower pore distribution in the range of 0.77 nm. The insertion of 10% Zn in the pillaring solution produced a material with a greater amount of Al in the pentacoordinate position and also presented better results of conversion (80%) and selectivity to ethylene (81%) in the ethanol dehydration reaction.

1. Introduction

Clays are natural materials composed of extremely small particles of one or more clay minerals and are very attractive scientifically and industrially. The versatility and variety of clays have been known since antiquity, but after being modified, they have gained even more technological importance due to their new properties, low cost, and abundance. In an attempt to obtain new materials that are more thermally stable and with greater specific area and porosity, clays have been subjected to modification processes such as, for example, pillarization [1,2]. Pillared clays are synthesized through the ion exchange of cations from the interlamellar region of the montmorillonite, such as Ca²⁺ and Na⁺, by polyoxocations of Al, Zr, Ti, and other metals. The resulting material after calcination contains oxides attached to the lamellae, functioning as pillars, keeping them separate, and giving rise to micropores [3].

The use of clays as catalysts is quite interesting, since they have characteristics such as high specific area, acidic sites, and micro- and mesopores, being fundamental for catalytic reactions. Acidity is an extremely important property of clays, influencing their catalytic performance in various processes, and depends mainly on the mineral and metals in the clay used for the synthesis of these materials [4]. Interest in pillared clays emerged in 1973 at the beginning of the oil crisis. At that time, new catalysts were sought that could crack heavy oil fractions, as the zeolites used had small pore sizes for such an application, which required pre-processing of the heavy fractions [5]. Initially, organic species were used for pillaring; however, they decompose at high temperatures, causing the clay structure to collapse [5]. Several metals such as Al, Fe, Zr, Cr, Ti, and Ga have been used since then, and in addition to being more thermally and hydrothermally stable, they can also provide catalytic properties to the material. Through the variation in parameters and methods of clay synthesis, we can control its structure and other textural properties according to the application.

Another possibility is mixed pillars, where another metal is incorporated into the formed pillar in addition to Al. Thus, in recent years, the applications of pillared clay have been diversifying with the incorporation of mixed pillars. For example, we see studies on the degradation of drugs such as paracetamol through a photo-Fenton process using clay pillared with Cu and Al [6] or a combination of polyethyleneimine to capture radioactive iodide from vapor [7]. Adding a metal such as cobalt to the pillarization process opens new perspectives for using pillared clays in widely known reactions, such as the Fischer–Tropsch process [8].

Among the different types of pillars found in the literature, few studies mention the use of pillars based on copper and zinc, which have catalytically interesting sites capable of acting in alcohol dehydration and dehydrogenation reactions. The replacement of aluminum using other metals in the Keggin ion structure is not always favorable or simple to identify. Frini et al. [9] present a study using different pillared methods. They used mixed solutions of Al and Cu, varying the percentage ratios of Cu/(Al + Cu) to 0, 5, and 10%. They observed a low incorporation of Cu, less than 1%. Mojović et al. [10] used Al/Cu pillared clays for toluene oxidation reactions with 1.6% CuO.

Some studies in the literature applied clays in the dehydration reaction of ethanol in ethylene. Ethylene is obtained on a large scale from the petrochemical industry, specifically from the cracking reaction of petroleum fractions. Compared to the traditional process, ethylene production from ethanol dehydration is economically viable and with greater purity [11], considering that it is part of a source from biomass. Jongsomjit et al. [12] applied montmorillonite clays (activated with mineral acids) and studied the conversion of ethanol and the selectivity to ethylene in a temperature range from 200 to 400 °C. In this study, conversion of 82% ethanol and 78% selectivity to ethylene was observed at a temperature of 400 °C. Xie et al. [8] studied PTMA-PILC clay modified by acid treatment, ion exchange, and impregnation, applied in the dehydration of ethanol in ethylene. The synthesis characteristics favored the catalytic performance in the sense of having more excellent activity for the dehydration of ethanol into ethylene, with the highest catalytic activity being 93.2% ethanol conversion and 99.1% ethylene selectivity at a temperature of 300 °C [13]. The clay was used as a support, where a diluted steam of bioethanol was applied in a dehydration reaction using heteropolyacid (dodecatungestophosphoric acid, DTPA) supported on clay. The highest conversion of 74% ethanol and 92% ethylene selectivity was observed with the 30% m/m DTPA/montmorillonite catalyst at a temperature of 250 °C compared to the other modified and unmodified catalysts studied. Chmielarz et al. [4] studied porous clay heterostructures (PCHs), obtained by the intercalation of silica, silica–alumina, silica–titania, and silica–zirconia pillars in the interlamellar space of montmorillonite. The best catalytic results in both studied processes were obtained from aluminum-doped SHPs, especially deposited using the matrix ion exchange method (TIE), which were more catalytically active in methanol for dehydration with dimethyl ether than γ-Al₂O₃, which is one of the most important commercial catalysts for this process. Although several studies have investigated different catalysts with acidic and basic sites for the dehydration and dehydrogenation of ethanol, the use of clays with copper and zinc pillars for application in ethanol conversion still needs to be explored to understand the action of active sites on the reaction process and confirm its viability.

In this study, pillared clays are prepared from mixed solutions containing Al and Cu and Al and Zn solutions and subsequently applied to ethanol conversion. The materials’ conversion efficiency and selectivity for ethylene are compared. The sequence of reactions involved and the participation of different active sites in the catalytic process are presented. Thus, this study seeks to understand the catalytic cycle and obtain essential data for developing new catalysts.

2. Materials and Methods

The starting material used was montmorillonite, provided by the Centro de Tecnologia Mineral (CETEM), called Bofe, but in this study is called montmorillonite. This clay is commercially used. It is extracted near the city of Boa Vista, PB, in the locality of Lajes. For the synthesis of the pillaring agent, the following reagents were used: Cu(NO₃)₂·8H₂O (Vetec, Duque de Caxias, Brazil) and ZnCl₂·2H₂O (Dinâmina, Indaiatuba, Brazil), and AlCl₃·6H₂O and NaOH (Sigma Aldrich, Darmstadt, Germany).

2.1. Preparation of Pillared Clays

The pillaring process was adapted from the literature [14]. Pillarizing solutions were prepared with 10% M (where M = Cu²⁺ or Zn²⁺) in relation to the total amount of solution with molar ratio OH/(M + Al) = 2.0. The pillarizing agent was prepared using the slow addition of a NaOH solution to a mixture of solutions of Al and M precursors (Cu or Zn) under continuous stirring, remaining under stirring at 60 °C for 24 h. We prepared the Al pillaring solution for comparison purposes.

After that, the pillaring agent was added to a suspension of previously expanded clay in the proportion of 1 g/100 mL of H₂O under constant agitation at room temperature for 2 h. Finally, the material was filtered until free of chloride ions indicated by testing with a silver nitrate solution and dried in an oven at 60 °C. Except for the copper-pillared clay, which was filtered with hot distilled water (80 °C) to remove the nitrate ions, the filtrate was monitored with UV-Vis. The materials were dried overnight in an oven at 60 °C and calcined at 450 °C for 3 h.

Pillarized clays containing different contents of Al, Cu, and Zn in the pillaring solution were synthesized, designated as Al100 (100% aluminum), Al/Cu10 (90% aluminum and 10% copper), Al/Zn10 (90% aluminum and 10% zinc), and Al/Zn20 (80% aluminum and 20% zinc). The percentages of copper or zinc are related to the pillaring solution.

2.2. Characterization Study

The samples were characterized using X-ray diffraction (XRD) in a diffractometer Bruker D2 Phaser using radiation CuKα (λ = 1.54 Å) with a filter of Ni, step size 0.02°, current of 10 mA, voltage de 30 kV, using a detector Lynxeye with divergent slit from 0.1 mm, time of 0.2 s, anti-air scattering screen of 1 mm and converging slit of 3 mm. The basal spacings of the samples were calculated using Bragg’s Law (nλ = 2dsinθ); n is an integer, λ is the wavelength of the incident radiation, d is the interplanetary distance, and θ is the angle of incidence.

The materials were characterized using X-ray fluorescence (XRF) on an appliance Bruker S2 Ranger using radiation Pd or Ag anode, max power 50 W, max voltage 50 kV, max current 2 mA, XFlash^® Silicon Drift Detector. From the results of the chemical analysis, values of AMR (Active Metal Ratio) were calculated from Equation (1), also presented below in Galeano et al. (2010) [15].

$$AMR\left(\%\right)=\left({\displaystyle \frac{mmol M}{mmol M + mmol Al}}\right)\times 100$$

(1)

Nitrogen adsorption analysis to 77 K were performed on an equipment ASAP 2020, and the pre-treatment was carried out to 60 °C for 12 h, then 200 °C overnight. For the analysis of the specific area, the method of Brunauer–Emmet–Teller (BET) [16] was used, to determine the volume of micropores and specific area of micropores, the α-plot was used, and to determine the total pore volume, the Gurvich method was applied at a partial pressure p/p₀ = 0.98.

Magic angle spinning nuclear magnetic resonance spectroscopy ²⁷Al (MAS – NMR) was processed on a device BRUKER AV400, Probe BL4 mm, Spinning speed: 10 kHz, frequency 104.26 MHz.

2.3. Catalysis Test in the Conversion of Ethanol

The test was carried out in a catalytic system at ambient pressure. The ethanol reagent was maintained at a temperature of 25 °C through a coupled thermostatic bath. In this sense, the reagent came into contact with the material deposited in a quartz fixed-bed reactor. Before the reaction, the material was subjected to a pre-treatment to clean traces of water and CO₂ at 300 °C for 30 min under a constant flow of N₂ (30 mL·min⁻¹). Then, the reaction was started, each test lasted 4 h under a continuous flow of N₂ (30 mL·min⁻¹) at 250 °C. The products were collected using a microsyringe with a capacity of 50 µL and injected into a gas chromatograph for identification and quantification. The gas chromatograph was from the brand Perkin Elmer, model Clarus 680, detector FID, column FS CAP Equity-5 (30 m × 0.25mm × 0.25 µm), bundled with poly (5% difenil/95% dimethyl siloxane). Ethanol conversion was calculated according to Equation (2), and ethylene selectivity was calculated according to Equation (3).

$$C_{et}={\displaystyle \frac{\left({Ethanol}_{in}-{Ethanol}_{out}\right)}{{Ethanol}_{in}}}$$

(2)

$$S_{prod}={\displaystyle \frac{{Ethene}_{out}}{n (hydrocarbon products)}}$$

(3)

The C_et is the ethanol conversion, Ethanol_in is the amount of ethanol before starting the reaction, Ethanol_out is the amount of ethanol collected at each point of the reaction, S_prod is the product selectivity, Ethene_out is the amount of Ethene generated and collected at each point of the reaction, and n (Hydrocarbon Products) is the total amount of product formed at each reaction point.

3. Results and Discussions

3.1. Pillared Clays with Mixed Al/Cu and Al/Zn Solution

Figure 1 shows the diffractograms of the montmorillonite sample and the samples intercalated with Keggin ion pillaring solution Al₁₃ (sample Al 100) and 10% of Cu or Zn (samples Al/Cu 10 and Al/Zn 10) and 20% of Zn (sample Al/Zn 20).

In the diffractograms of all intercalated samples, it is possible to observe a displacement of the reflection referring to the (001) plane, indicating an increase in the basal spacing. The calculated basal spacing for montmorillonite was 1.5 nm. According to Gomes (1988), when the interlamellar cation is Ca²⁺, the spacing is approximately 1.55 nm. Diffraction analysis also reveals different phases, such as smectite, quartz, and cristobalite, common in montmorillonite minerals and which may be due to weathering [17]. The results obtained for the intercalated montmorillonite samples showed that the pillaring process increased the basal spacing of the montmorillonite from 1.6 nm to 1.88 nm for the samples pillared with Al/Cu and Al/Zn. The incorporation of copper and zinc in the pillaring solution did not promote a difference in the values of basal spacing when compared to montmorillonite intercalated only with aluminum. This indicates that the intercalation process took place effectively and the basal spacing distances were not influenced by the addition of Cu and Zn. The calcination process is responsible for the stabilization of the structure due to the micropores originated by the oxide pillars present, which will be confirmed through the analysis of the N₂ adsorption–desorption isotherms. In this study, the calcination parameters were already well defined for pillared clays with aluminum, described by [18]. The calcination procedure consisted of calcining the pillared samples at 450 °C for 3 h following a heating rate of 5 °C/min. Figure 2 shows the diffractograms of the calcined and pillared samples with Al/Cu and Al/Zn and for the montmorillonite sample.

In the XRD patterns presented in Figure 1, we can observe that, for the Al/Zn10 sample, a reflection of around 9° (2 theta) is observed. This reflection indicates that some lamellae are collapsed and not intercalated [3]. It can be observed in the XRD patterns of Figure 2 that, due to the pillarization process, the reflection related to the (001) plane remains after calcination, and the reflection due to the collapsed lamellae is not very intense or is almost imperceptible. The interlamellar spacing calculated for the samples with calcined Al/Cu and Al/Zn pillars were around 1.6 and 1.8 nm, respectively. The difference between the intercalated and calcined samples’ values is due to the dehydroxylation process and the formation of oxides that now act as pillars [19]. The calcined sample presented structural collapse, evidenced by the value corresponding to the 001 reflection, which measured 0.96 nm and became almost imperceptible. According to Brown and Brindley (1980) [20], the minimum distance between the lamellae for the smectite clay mineral, at approximately 2θ = 9°, is 0.96 nm.

After the pillarization process, the quartz and cristobalite phases are still observed, indicating that the insertion of pillars affects only the montmorillonite phase and not the other phases present.

XRF analysis (

Table 1. Percent chemical composition obtained using FRX for all pillared samples and calcined montmorillonite, and the Active Metal Ratio (AMR%) calculated for the calcined montmorillonite, Al100C, Al/Cu10C, Al/Zn10C, and Al/Zn20C samples.

	Calcined Montmorillonite	Al100C	Al/Cu10C	Al/Zn10C	Al/Zn20C
Na2O	0.00	0.35	0.36	1.61	1.07
MgO	2.90	2.12	2.22	2.61	2.08
Al2O3	16.03	19.09	19.33	20.97	22.51
SO3	0.10	0.06	0.06	0.30	0.33
Cl	0.11	0.09	0.08	0.09	0.09
K2O	0.36	0.25	0.24	0.37	0.29
CaO	1.24	0.00	0.00	0.09	0.08
TiO2	0.93	0.84	0.80	0.12	0.09
Fe2O3	8.12	7.08	6.73	3.60	2.89
CuO	0.00	0.00	0.06	0.00	0.00
ZnO	0.00	0.00	0.00	0.12	0.44
AMR (%) Pillared Clay			1.15	1.50	4.08

For all the samples or percentages of SiO₂ considered to be 70.20%, equal to the SiO₂, the theory gives the starting sample.

) shows that the amount of Cu and Zn incorporated through pillarization was very low. Therefore, no reflections related to these phases are observed in the diffractograms presented.

From the results of the chemical analysis (Table 1), AMR (Active Metal Ratio) values were calculated, as shown in Table 1. The AMR value is defined as the atomic percentage ratio between the active metals (Cu or Zn) and the total metal content (Al + Cu or Zn) in the pillaring solution. Galeano et al. (2010) [15] calculated the AMR value from the pillaring agent. In order to better compare the changes that occurred using the treatments in the chemical composition of the solids, the Al₂O₃, CuO, or ZnO contents were normalized to the SiO₂ content in the starting montmorillonite (70.20%), assuming that the SiO₂ content is not affected by the treatments performed. The AMR% values for the pillared clays with Cu and Zn insertions are presented below in Table 1.

According to Galeano et al. (2010) [15], 2% Cu values improve the catalytic properties of the materials in peroxide catalytic oxidation (CWPO) reactions of the methyl orange dye, and above 2% (product AMR = 4.23), it was observed that the activity decreases. In the present study, the calculated AMR value was lower than that obtained in the cited study. It was expected that Cu would be incorporated into the pillars; however, this could be hampered due to the difference between the sizes of the ionic radii of the metals (Cu²⁺ 0.72 Å, Zn²⁺ 0.74 Å, and Al³⁺ 0.54 Å), which can make an isomorphic substitution in the Keggin ion impossible [15]. The low capacity of copper to stabilize in clays pillared by conventional pillaring procedures has been previously reported [21,22]. From the calculations, the sample with 20% zinc content was the one with the highest AMR % value (4.08), which indicates the greater ease of zinc incorporation.

Table 2. Texture analysis data for the following samples: montmorillonite, Al100C, Al/Cu10C, Al/Zn10C, Al/Zn20C.

Sample	BET	α-Plot		Vt
	SBET (m2/g)	St (m2/g)	Vo (cm3/g)	VTP (cm3/g)
Montmorillonite	123	56	0.04	0.19
Al100C	198	48	0.07	0.19
Al/Cu10C	193	72	0.06	0.19
Al/Zn10C	208	71	0.07	0.19
Al/Zn20C	241	103	0.08	0.21

S_BET—specific area, S_t—specific external area, V_o—micropore volume (α-Plot), Vt—total pore volume (Gurvich).

presents the textural properties obtained from the results of the N₂ adsorption–desorption analysis. To verify the characteristics of the samples, some mathematical methods were applied to analyze the data present in the isotherms, choosing a certain range of p/p0.

In addition to increasing the basal spacing, pillaring also promotes an increase in the specific area and modifies the type of porosity of the material. The montmorillonite already has a significant specific area, which can be related to the spaces between particles and also the aggregates of particles [23]. Pillarized samples had larger specific areas when compared to starting clay. The material has a more uniform porosity due to the creation of micropores between the lamellae. An increase in micropore volume from 0.04 cm³/g to 0.07–0.08 cm³/g is observed. For the samples Al100C, Al/Cu10C, and Al/Zn10C, the values of the textural properties are close, which indicates that the insertion process of the pillars is not being affected by the decrease in Al and the addition of Cu or Zn (in the synthesized amounts) in the pillars’ solutions. For the Al/Zn20C sample, there is a significant increase in the external specific area, which may indicate a deposition of ZnO particles under the surface of the slides.

The N₂ sorption isotherms of montmorillonite and pillared samples calcined at 450 °C are shown in Figure 3a, and the pore size distribution is shown in Figure 3b.

All isotherms are type IV, and low pressures have a type I isotherm profile (nomenclature IUPAC) [24], characteristic of materials that present microporosity, which can be confirmed at low relative pressure points, Figure 3a. Hysteresis can be classified as type H3, which indicates the presence of slit-like pores, characteristic of clays. The specific area calculated from the BET equation for montmorillonite already presents a high value compared to other clays of the bentonite type. With pillarization, these specific area values increased by about 36–49% in relation to the starting montmorillonite. This increase in the specific area has a direct contribution from the increase in the volume of micropores. Another observation is the greater homogenization of the size of the micropores present, which is observed by the mean size distribution of the micropores. For all samples, the mean micropore size was estimated at 0.77 nm, as shown in the pore diameter distribution in Figure 3b. This value (0.77 nm), added to the lamella thickness (0.96), gives 1.73 nm, which confirms the results presented with XRD where the pillared samples had a basal spacing between 1.6 and 1.8 nm.

Figure 4 shows the NMR spectra of Al. The resonances of Al^VI (2 to 3.5 ppm), Al^V (approximately 30 ppm), and Al^IV (between 51 and 67 ppm) for pillared clays are highlighted [25,26].

In montmorillonite, from the deconvolution of the signals, the presence of two signals can be observed with displacements of 51.1 and 67 ppm attributed to the presence of tetrahedral Al. The deconvolution of the band into two signals indicates the presence of crystallographically non-equivalent sites for tetrahedral aluminum [27]. The presence of these signals remains in the samples even after the pillaring process. In

Table 3. Relations of signal areas for Al with different coordinates: Al^VI/Al^IV e Al^IV/Al^V for natural and pillared clay.

	Natural Clay	Al100C	Al/Cu10C	Al/Zn10C	Al/Zn20C
AlVI/AlIV	9.9	7.7	9.2	9.1	7.3
AlIV/AlV	nd	5.2	4.8	2.1	4.0

, the relationships between the areas of the signs referring to Al^IV, Al^V, and Al^VI are calculated from the spectra of NMR.

Using the obtained area values and the deconvolution of the NMR spectrum signals, it is observed that the relation of Al^VI/Al^IV in clay pillared with Al decreases compared to montmorillonite. The Keggin ion is composed of a central tetrahedral aluminum (Al^IV) surrounded by 12 octahedrally coordinated aluminum atoms (12 Al^VI) [28]. Therefore, the insertion of these species justifies the increase in the amount of Al^IV and the consequent decrease in this Al^VI/Al^IV ratio. However, the calcination process at 450 °C favors the formation of pentacoordinate aluminum, with the loss of terminal groups binding the H₂O of octacoordinated aluminum [29]. These pentacoordinate aluminum sites transform into strong Lewis acid centers [29].

In pillared materials with the insertion of Cu and Zn, there is a decrease in the Al^VI/Al^IV ratio compared to montmorillonite due to the insertion of Al₁₃ pillars, as well as the appearance of pen coordinated by the calcination process. When comparing the samples with each other, the Al/Zn10C sample has the highest aluminum content with a pentacoordinate; however, the increase in the Zn content does not have the same effect, suggesting that the Zn content can influence the acidic properties and, consequently, the catalytic properties. In a recent study by Oliveira et al. (2020) [30], they observed that the impregnation of Zn (2% by mass) in a pillared clay increased the Lewis-type acidity in these materials. Higher Zn contents (20%) may cause the formation of ZnO aggregates, thus decreasing the interaction with the Al of the columns.

Comparing the samples with 10% Cu and 10% Zn, a similar Al^VI/Al^IV ratio is observed, but with different Al^VI/A^lV ratios, indicating that Zn interacts more with the pillars, forming a greater amount of pentacoordinate Al.

3.2. Catalysis Test in the Conversion of Ethanol

The clays were submitted to ethanol conversion as a model reaction to evaluate the acid–base properties of the different synthesized materials and to understand the role of active sites in ethanol dehydration and dehydrogenation. The different results obtained are shown in

Table 4. Conversion average and selectivity for different clays.

Sample	Conversion (%)	Ethylene (%)	Diethyl Ether (%)	Ethanal (%)
Montmorillonite	8	27	70	3
Al100C	36	42	54	4
Al/Zn10C	80	81	11	8
Al/Zn20C	74	60	28	12
Al/Cu10C	46	54	41	5

.

Through the products obtained, it was possible to propose the main reactions involved in the catalytic process, described below.

$${CH}_3{CH}_2{OH}_{\left(g\right)} \leftrightarrow {CH}_2={CH}_2+H_2{O}_{\left(g\right)}$$

(4)

$$2{CH}_3{CH}_2{OH}_{\left(g\right)} \leftrightarrow (C_2{H}_5{)}_2{O}_{\left(g\right)}+H_2{O}_{\left(g\right)}$$

(5)

$${CH}_3{CH}_2{OH}_{\left(g\right)} \leftrightarrow {CH}_3{CHO}_{\left(g\right)}+H_{2\left(g\right)}$$

(6)

It is observed that reaction (4) occurs because of the dehydration of ethanol in ethylene, producing water. In addition, other parallel reactions were identified, such as the dehydration of ethanol in diethyl ether, also producing water (5) and, in a smaller amount, ethanal and H_2(g) through the dehydrogenation of ethanol (6). Figure 5a shows the conversion/selectivity as a function of time for untreated montmorillonite.

The montmorillonite sample (Figure 5a) showed lower ethanol conversion; moreover, it was more selective in the production of diethyl ether (reaching 75%), also forming ethylene (maximum of 31%), and small quantities of ethanal. Montmorillonite-type clays in their natural form may have both weak Bronsted and Lewis acidity. Bronsted acidity can be attributed both to the presence of compensation cations between the laminae and the presence of terminal OH groups at the edges of the layers, and, in this case, Lewis acidity to the presence of coordinately unsaturated Al also present at the edges of the lamina [31], as shown in the NMR results (Figure 5).

The pillaring process increases the Lewis acidity of the clays. Figure 5b shows a graph of the conversion/selectivity as a function of time for a clay pillared with Al₁₃. The Al100C sample presents higher ethanol conversion values compared to montmorillonite, in addition to greater selectivity for ethylene formation and a decrease in the formation of diethyl ether. Observing these results and correlating them with table III, the data indicate the influence of pillarization on catalytic performance. This clay has greater microporosity compared to montmorillonite and greater homogeneity in micropore size, according to N₂ isotherms. Furthermore, it also has Al in a pentacoordinate position, according to the NMR results. The presence of this type of Al is reported as reversible in the hydration of materials, justifying its better performance.

The insertion of Cu from the pillaring solution produced a material with textural characteristics very similar to pillared clay with Al alone, as well as producing additional active sites capable of converting ethanol. In Figure 5c, one can see the effect of adding copper to Al/Cu10C clay concerning conversion and selectivity.

The Al/Cu10 sample showed an increase in conversion due to the presence of additional Cu sites compared to Al100C clay. However, the increase in selectivity for ethylene over diethyl ether was more significant, producing a smaller amount of ethanal. The selectivity variation is directly related to competition between the two main sites of Al and Cu in the interaction with ethanol. It is known that CuO sites are related to dehydrogenation reactions [32], whereas Al₂O₃ sites are related to dehydration reactions [33]. However, according to the FRX results, the addition of CuO in the material is less than 0.1%, influencing very little in preferential terms of the route involving the dehydrogenation reactions, justifying the low selectivity for ethanal. Material with copper has a lower amount of pentacoordinate Al compared to samples containing Zn according to Table 3, and, consequently, has a lower Lewis acidity, leading to lower performance in the conversion of ethanol. The deposition of coke on the catalyst surface also leads to a decrease in conversion during the first minutes of reaction, preventing the activity of the few Cu sites present and favoring parallel reactions [33].

Among the series of clays studied, the materials Al/Zn10C (Figure 5d) and Al/Zn20C (Figure 5e) showed the highest percentage of conversion, as observed in Table 4. Thus, the Al/Zn10C sample was the most selective in the production of ethylene, as by-products of the formation of ethanal and diethyl ether occurred in smaller amounts.

The Al/Zn20C sample showed high ethanol conversion compared to other studied clays, as well as being more stable over the 4 h of reaction. The material presented relatively high selectivity values in the ethylene formation, reaching 64% in the 35 min reaction and 54% at 222 min. There was also the formation of diethyl ether (38% in 222 min) and ethanal in smaller amounts (9% in 222 min).

As for the textural properties, the Al/Zn20C sample had a larger external area, which may indicate the formation of ZnO particles on the surface of the lamellae. The presence of ZnO incorporated in pillared clays promoted an increase in Lewis acidity, as shown by the NMR results, justifying its catalytic performance.

The catalytic performance of clays during the reaction decreases naturally, as the active sites are obstructed due to the formation of coke. The Al/Zn10C clay showed greater selectivity for ethylene production, while the others showed lower selectivity. Regarding ethanol conversion, the Al/Zn10C and Al/Zn20C clays presented better performance due to their higher acidity, while the montmorillonite indicated low conversion due to its smaller area and the low exposure of the sites. In samples containing zinc, the solid Al/Zn10C showed better catalytic performance due to a greater amount of pentacoordinate Al and, consequently, better acidity. Therefore, the results show that the type of site and exposure of active centers directly influence the performance for ethanol conversion.

3.3. Mechanism Proposal Conversion of Ethanol Use Pillared Clay

From the discussed results of the ethanol dehydration/dehydrogenation reaction at the different active sites based on the different synthesized solids, it is possible to propose a reaction mechanism. Figure 6 shows the three main routes: dehydration obtaining ethylene (I), dehydration generating diethyl ether (II), and dehydrogenation to ethanal (III). The five synthesized materials containing Al, Cu, and/or Zn were represented on the left of Figure 6; however, the acid–base sites were generated as M-O-M-O according to the illustration on the right, where M represents the Lewis acid centers and O represents the Lewis acid centers from the network oxygens, considering that the different prepared catalysts have these types of sites. It is noteworthy that the reaction environment is complex, since clay is composed of several metals, as indicated in the FRX analysis.

It is known that the selectivity for ethylene formation is temperature-dependent since it is an endothermic reaction (+44.9 kJ/mol) and is favored at higher temperatures, while the selectivity for diethyl ether formation is exothermic nature (−25.1 kJ/mol) and is favored at lower temperatures. Despite being two competitive routes, another factor besides temperature that can directly influence the selectivity is the presence of Lewis acid sites, whether weak, strong, or moderate [34].

The first step of the catalytic cycle is the preferential adsorption of ethanol on the acidic and basic Lewis sites. Thus, in route (I), after adsorption of the reagent on clay, the oxygen from the ethanol interacts with the Lewis acid site and the hydrogen from the hydroxyl group with the oxygen from the lattice interacts with the basic site. After the interactions, the removal of 2H and O²⁻ occurs via active sites on the clay and the reorganization of electrons. At the end of the route, ethylene and H₂O are obtained, which are desorbed as products, as shown in reaction (1).

In route (II), the adsorption of two ethanol molecules occurs at close sites, from the removal of hydroxyl from one molecule and dehydrogenation from another. Dehydration leads to the formation of H₂O and with electronic rearrangement, diethyl ether is formed, which are desorbed, completing the catalytic cycle according to the reaction (two). This route has been investigated in theoretical studies (Austin et al., 2018), indicating that ethanol molecules react through a combined transition state, in which the diethyl ether and H₂O products are formed in a single step. It is worth emphasizing that in route (II), in dehydration, the ethanol molecule interacts with Lewis acid (Mⁿ⁺) and basic (O²⁻, lattice oxygen) sites, while in dehydrogenation, the ethanol molecule interacts only with basic sites and oxygen from the network.

Furthermore, the literature proposes another route where diethyl ether is formed from an ethylene molecule interacting with an ethanol molecule under Lewis acid sites [35].

In route (III), after adsorption of the ethanol molecule on the acidic and basic Lewis sites (similar to route I), a first dehydrogenation occurs followed by an electronic rearrangement. Finally, another dehydrogenation reaction takes place, obtaining the ethanal and H₂ molecules, which are desorbed as products according to the reaction (3).

Thus, observing the data obtained from the reactions, it is possible to indicate that the metal support synergistic effect was more efficient for the samples containing the interaction of the Al/Zn sites compared to the Al and Al/Cu sites, showing that the synergistic action between these two Lewis acid type sites led to higher values of conversion, selectivity, and stability, in addition to the effect of pillaring treatment when compared to montmorillonite.

Finally, it is worth emphasizing that previous studies have already investigated the dehydration and dehydrogenation of ethanol in the presence of catalysts based on aluminum, copper, and/or zinc oxides [34,36,37]; on the other hand, the results presented in this work confirm that the combination of copper and zinc oxides with pillared clay present promising catalytic results compared to previously published studies using similar oxides, but in a different structure from clay [38,39,40].

4. Conclusions

The montmorillonite-type clays were pillared with mixed solutions of Al/Cu and Al/Zn, and all samples showed basal spacing consistent with pillared samples as well as increased specific area and micropore volume. The percentage of active metal ratio (AMR) for the samples with 10% Cu and Zn were very close, as were the values calculated from the NMR data for the octahedral/tetrahedral aluminum ratios. In the case of the Al/Zn20 sample with 20% Zn, the NMR values were the highest, as well as the values of specific area, external area, and total volume and micropores. Regarding the octahedral/tetrahedral aluminum ratios, this sample was the one that most resembled the material Al100 pillared with aluminum.

According to the characterizations carried out, the presence of Cu and Zn did not interfere in the pillaring process due to the size of these species. Thus, it can be deduced that Cu and Zn decorate the pillars, leave the Al pentacoordinate, and, consequently, are more active in ethanol conversion due to the greater exposure of these sites in the catalytic reaction.

The Al/Zn10 and Al/Zn20 samples showed higher percentages of ethanol conversion compared to other materials with Cu-Ale; the Al/Zn10 catalyst exhibited greater selectivity for ethylene. However, both samples differed in relation to the aluminum tetrahedral/pentacoordinate ratio, and the Al/Zn10 sample was the one with the highest amount of pentacoordinate Al, justifying its higher selectivity to ethylene due to its higher Lewis acidity. Therefore, the exposure and nature of acidic sites directly influence the catalytic performance.