Investigation of NH₃ Desorption Kinetics on the LTA and SOD Zeolite Membranes

Abstract

The acidity characteristics of zeolite are highly significant, and understanding the acidic properties is essential for developing new types of zeolite catalysts. Zeolite membranes were synthesized using metakaolin, sodium hydroxide, and alumina with a molar ratio of 6Al₂Si₂O₇:12NaOH:2Al₂O₃ as the starting ingredients. X-ray diffraction, scanning electron microscopy, and infrared spectroscopy were used for this study. N₂ adsorption measurements determined the surface areas of the SOD zeolite membrane (115 m²/g) and the LTA membrane (150 m²/g). The units of absorbed water vapor were 40 and 60 wt% for the SOD membrane and the LTA membrane, respectively. The strength and number of acid sites of the synthesized LTA and SOD zeolite membranes were determined by temperature-programmed desorption of ammonia. As a result, the value of the total acidity of the LTA zeolite membrane is in the range of 0.08 × 10¹⁹ units/m² while that of the sodalite membrane is an order of magnitude lower and is 0.006 × 10¹⁹ units/m². The apparent activation energy values for desorption of ammonia from LTA and SOD zeolite membranes were calculated using data on the kinetics of desorption of ammonia at different heating rates. It was found that at temperatures below 250 °C, the degree of conversion of the activation energy values is no more than 35 kJ/mol, which corresponds to the desorption of physically bound ammonia. An increase in the activation values up to 70 kJ/mol (for SOD) and up to 80 kJ/mol (for LTA) is associated with the desorption of chemically bound ammonia from the samples.

1. Introduction

Zeolites and membranes based on them are widely used in various separation processes of mixtures and gases in pressure swing adsorption (PSA) processes [1,2]. The use of zeolite membranes in PSA processes is determined by their capacity to separate at the molecular level [3]. The study of the economic feasibility of PSA on zeolites when obtaining oxygen in air separation units showed that the energy intensity of this process is significantly lower than in liquid air rectification units of the same efficiency [4,5]. However, under the conditions of PSA processes, zeolites tend to exhibit variability in capacitance characteristics and a strong dependence of the adsorbent properties on the preparation, storage, and use [6,7,8,9]. Therefore, one of the main indicators is the properties of the surface of the adsorbent, which play an essential role in predicting their use in PSA processes.

The developed acidic properties of zeolites arise due to the inclusion of an aluminum atom in the tetrahedral region of the framework; the acidity depends on the zeolite framework microstructure, including the presence, number, and size of cavities. Zeolite membranes are among the most promising materials used on an industrial scale due to their unique properties, such as a high uniformity and pore size and high thermal and chemical stability. Zeolite membranes are the only membranes capable of performing separation at the molecular level [10,11]. In this case, the study of zeolite membranes is carried out for further prediction and use in low-cycle adsorption processes, and for gas separation and pervaporation. The factors determining the use of zeolite membranes in these processes include the acid-base properties of the surface and the structural features of the assembly of the zeolite framework.

Among the variety of methods used for measuring surface acidity, due to the availability, speed, and good visualization of results, the most popular method is temperature-programmed desorption (TPD) [11,12,13]. The use of TPD makes it possible to record the energy spectrum of samples by desorption of probe substances, and to quantify the concentrations of active centers and the activation energies of desorption of probe substances in various forms of desorption [11,12,13].

Thus, the purpose of this study was to investigate the surface properties of materials based on zeolites and their further use in pressure swing adsorption processes.

2. Materials and Methods

Powder X-ray diffraction (XRD) patterns were recorded on a Bruker D8 X-ray diffractometer (Bruker, Billerica, MA, USA). Copper Kα radiation (λ = 0.15406 nm, Ni–filter) was used with a 40 kV and 20 mA power supply. The scan rate was 1 min⁻¹, and the scanning step was 0.01. The crystalline phases in the XRD patterns were identified by comparison of the calculated interplanar spacings (d = λ/2 sin Θ, where λ is the wavelength and Θ is the diffraction angle) with those taken from the ASTM database.

Since the LTA and SOD zeolites used in this study have a cubic structure, the zeolite lattice parameters were calculated as:

$$\frac{1}{d^2}=\frac{h^2+k^2+l^2}{a^2}$$

(1)

where h, k, and l are Miller indexes.

Broadening of the X-ray diffraction profile allows one to determine both the dimension of the coherent scattering region (CSR) and the value of mean-square micro deformations (MDs). For this purpose, we used the modified Scherrer’s equation [14,15]. The modified Scherrer’s equation (so-called Scherrer–Selyakov equation) can be written as:

$${\beta }_{ph}=\frac{\lambda }{D_{CSR}\cos \Theta }+4\epsilon \tan \Theta$$

(2)

or in a linear form:

$${\beta }_{ph}\cos \Theta =\frac{\lambda }{D_{CSR}}+4\epsilon \tan \Theta$$

(3)

where β_ph is the physical component of the broadening, DCSR is the CSR dimension, ε is the mean-square MD value, and Θ is the position of the profile centroid of the sample. The value of β_ph can be extracted from the total broadening profile using a Gaussian distribution as:

$${\beta }_s^2={\beta }_p{{}_h}^2+{\beta }_s{{}_t}^2$$

(4)

where β_s is the integral half-width of the sample profile and β_st is the integrated half-width of the standard sample. It was assumed that for the standard sample (Ecolab, Moscow, Russia), the measured broadening is equal to the instrumental broadening only and is associated with both the device characteristics and exposure conditions.

The Fourier transformed infrared (IR) spectra were measured using an Avatar 360 FT–IR ESP spectrometer (Thermo Fisher Scientific, Waltham, MA, USA) in the range of 4000–400 cm⁻¹. The samples were prepared using the KBr method with a sample-to-KBr ratio of 1:100.

In this study, ammonia was used as a probe to investigate the acid-base properties of zeolite structures. The choice of ammonia is due to its high degree of basicity, which makes it possible to determine strongly acidic centers and weak centers by the small size of the given molecule [11,13].

The concentration of acid sites in the investigated samples was identified by the units of ammonia desorbed at the moment of fixation of the desorption peaks. The accuracy of determining the units of desorbed ammonia by gas chromatography was ±5%.

Temperature-programmed desorption of NH₃ was performed using a Setaram DSC 111 differential scanning calorimeter (Caluire, France) consisting of a flow measurement and switching system and a cylindrical oven controlled by a linear temperature programmer (Omega CN 2010). A mass spectrometer (MS, Thermostar from Pfeiffer, Aßlar, Germany) was used as a detector.

The adsorption of NH₃ estimated the total surface acidity of the samples by the number of chemisorbed molecules. The heating rate was 10, 15, 20, 25, and 30 °C/min. The following formula determined the number of active centers:

$$N_i=\frac{6\cdot {10}^{23}{}^{}\cdot S\left(T_{maxi}\right)\cdot V}{\mathrm{22,400}\cdot S_s{}_p\cdot {\displaystyle \sum }S\left(T_m{}_{axj}\right)\cdot m}$$

(5)

where 6.022 × 10²³ is Avogadro’s number; S(T_maxi) is the area under the corresponding maximum on the thermal desorption curve, mm²; S_sp is the specific surface of the support samples, m²/g; G is the weight, g; and ΣS (T_maxi) is the total area of peaks on the thermal desorption curve, mm²; mL. The number of moles of desorbed ammonia was calculated by V/22,400, where V is the desorbed volume of ammonia (mL, NTD).

Samples for TPD were loaded into a quartz glass tubular reactor, and the resulting layer was fixed with quartz wool. Before taking the TPD values of ammonia or carbon dioxide, the sample was heated to 650 °C for 60 min in a He flow (30 mL/min) to remove adsorbed components, such as H₂O. The adjusted pretreatment temperature provided reproducible TPD conditions and avoided sintering effects. Then, the sample was cooled to 50 °C. The adsorption of ammonia was carried out from an ammonia-gel mixture with a concentration of 10 vol.% NH₃. After saturation, it was washed with He.

N₂ adsorption-desorption isotherms were measured at 77 K on a Sorbi–MS analyzer (Novosibirsk, Russia). Samples were outgassed at 573 K before the adsorption measurements. The specific surface area was calculated from nitrogen adsorption data in the relative pressure range from 0.05 to 0.2 using the BET (Brunauer–Emmett–Teller) equation. The total pore volume was estimated from nitrogen adsorbed at a relative pressure of about 0.99.

The scanning electron microscopy (SEM) measurements were taken with a JSM–6460 LV microscope (JEOL, Ltd., Tokyo, Japan).

To synthesize the zeolite membrane, we used kaolin for the perfume industry (Prosko Resursy, Dnipro, Ukraine). The content of the kaolinite phase was 97.7 wt%. Metakaolin, in the form of an amorphous fine-dispersed white powder, was prepared by the calcination of kaolin at 700 °C for 4 h. Moreover, commercial sodium hydroxide (Kaustik, Volgograd, Russia) in the form of flakes was used. The NaOH content was 99.5 wt%. Commercial aluminum hydroxide (SUAl, Shelekhov, Russia) was used. The aluminum hydroxide contained 98.8 wt% gibbsite. Aluminum hydroxide was converted to γ-Al₂O₃ as an amorphous fluffy white powder during the calcination process at 550 °C for 4 h.

To synthesize membranes, amorphous matrices based on porous anodic alumina separated from an aluminum support (10 × 10 nm matrix, 20 nm diameter, 20 μm thickness; manufactured by Nelan-Oxide Plus, Petrozavodsk, Russia) were used as a support.

In the first step, to fix the components on the support surface, the support was immersed in the initial suspension and subjected to USP (ultrasonic-assisted processing) for 10 min. The molar ratio of the mixture ingredients was Al₂Si₂O₇/NaOH/γ-Al₂O₃ = 6: 12: 2. This ratio was selected to ensure sodium aluminate formation in the first step of LTA and SOD zeolite membrane synthesis. For this purpose, excess γ-Al₂O₃ was additionally introduced into the initial reagent mixture. The optimal γ-Al₂O₃ excess is 2 mol alumina per 6 mol metakaolin (33.3 mol%). Suspensions with a mass weight ratio of solid to liquid (S/L) of 5: 1 (with distilled water as a solvent) were prepared from the initial mixture.

Zeolite membranes were synthesized by solution crystallization without seed crystals. The crystallization step utilized a 2 mol/L and 8 mol/L NaOH solution in appropriate units to ensure S/L = 10: 1 (volume ratio). Al₂O₃ supports with a pre-applied layer of sodium aluminate/aluminosilicate precursors prepared in the previous USP step were dried at 60–80 °C and calcined at 650 °C before they were immersed in the solution. Crystallization was carried out with two solutions by conventional heating in a 2 and 8 mol/L NaOH solution (pH 14) at 80–90 °C for 2 h, for the synthesis of the LTA and SOD zeolite membrane, respectively.

3. Results and Discussion

The main factor determining the selectivity of zeolites and the main distinguishing feature from other materials is the shape and size of the pores and cavities in their structure. Considering membranes based on LTA and SOD zeolites, the optimal parameters for their production were defined in [8,9].

The main characteristics of the obtained membranes are shown in Figure 1 and Figure 2. We can see that the pore size of the crystal lattice a of the obtained samples is close to theoretical values and corresponds to SOD = 8.88 Å, LTA = 24.76 Å. The crystal lattice deformation level is insignificant and does not exceed 0.15%. The calculated values of the coherent scattering region (Dcsr) are 462 nm for SOD (Table in Figure 1) and 780 nm for LTA (Table in Figure 2), which is natural and takes into account the differences in the arrangement of the sodalite structures of these samples. In the case of sodalite, cuboctahedra are linked through simple four-membered rings (Figure 1e), whereas with LTA zeolite, they are linked through double four-membered rings (Figure 2e). The cuboctahedra link forms a system of regular large α-cavities with a diameter of 1.1 nm (LTA zeolites), connected by rings 0.4–0.5 nm in diameter. Such rings are spaces that open access to the volume in which molecules are adsorbed. Similar to amorphous adsorbents, rings are identified with pores. LTA zeolite is characterized by small micropores 0.3–0.45 nm in size [14].

The total surface acidity of the samples was estimated from the adsorption of NH₃ and the number of chemisorbed molecules, based on the assumption that each probe molecule occupies one acid-base center on the surface [13,15,16,17,18,19]. At the same time, regardless of the type of membranes, gas desorption was fully completed with the temperature increase in the reactor to 600 °C. In general, if the desorption of physically adsorbed molecules does not require thermal activation, it is heating that contributes to the beginning of the desorption of chemisorbed probe molecules. Desorption takes place from the same centers in a specific temperature range, and, consequently, the typical maximum temperature characterizes the bond strength (acid-base strength of the surface center).

When identifying acid sites on the surface of metal-containing samples by the TPD of ammonia, the choice of a temperature of 200 ± 50 °C was proposed as a conventional boundary [13]. Thus, below this value, desorption of ammonia is associated with weak Bronsted centers in the form of OH-NH₃ adducts (surf). At a higher temperature, molecules held on the surface by stronger (Lewis) acid centers are desorbed. Lewis acid and basic centers are coordinatively unsaturated surface cations and anions, respectively. This phenomenon is characterized by vacancies in the coordination sphere and a strong covalent bond with an NH₃ molecule [17,18,19]. At the same time, the acid sites in zeolites cannot be explained only by the presence of hydroxyl groups. An investigation into the adsorption of pyridine on zeolite, shown in [20], revealed that after heating the sample with adsorbed pyridine at 200 °C, the formation of pyridinium ions was observed in the spectra, which indicates the presence of Bronsted centers in the zeolite. It is not clear whether these centers are formed during the rearrangement of the zeolite framework and the migration of traces of water in the zeolite under calcination or whether they are formed as a result of decomposition of the adsorbed pyridine. A schematic representation of the process of desorption of ammonia is shown in Figure 3.

It was found that the TPD spectra of ammonia desorbed from the surface of the sodalite membrane (Figure 4) exhibit four desorption temperature maxima (T₁–T₄). They can be conventionally ranked as weakly acidic (T₁, T₂) and moderately acidic (T₃), and strong (T₄). In this case, the desorption of ammonia was completely accomplished when the temperature in the reactor increased to 500 °C. From the data of the TPD spectra recorded at different heating rates, it can be clearly seen that the heating rate does not significantly affect the position of the peaks. The low-temperature region up to ~200 °C is characteristic of the desorption of physically adsorbed ammonia molecules adsorbed on unsubstituted cationic sites (100–120 °C), and the desorption of ammonia from weak acid sites (150–170 °C) [18,19,21]. The high-temperature region at 250–500 °C corresponds to the desorption of ammonia adsorbed on medium and strong acid sites.

The TPD spectra of ammonia desorbed from the surface of an LTA zeolite membrane (Figure 5) differ from the TPD spectra on a sodalite membrane and have five temperature maxima. T₁, T₂, and T₃ also belong to the low-temperature region, which is characteristic of the desorption of physically adsorbed NH₃ molecules [22]. Ammonia is adsorbed on acid sites, which most likely have electron-withdrawing properties, based on the structural features of LTA zeolite (Na₁₂Al₂Si₁₂O₄₈), which are tetrahedral compounds. Thus, desorption on the LTA membrane in the region of weak acid sites is more complex (multistage). The intensity of the high-temperature peak during desorption on the LTA membrane is higher than that of sodalite, but the peak in the region of medium acidity is, on the contrary, less intense. In addition, with an increase in the crystal size, the maximum of the high-temperature peak shifts from 360 to 400–430 °C. This is explained by the fact that, under the condition of possible desorption of ammonia after removal from the acid site, the position of the peak maximum is influenced by both the structural features of the pores and the total number of acid sites [18,19,21,22]. Therefore, with a more significant number of acid sites of the same type, the maximum of the corresponding peak shifts towards higher temperatures.

From the obtained TPD spectra recorded on both types of membranes, it is possible to draw some conclusions about the effect of the heating rate; namely, at a higher heating rate, 30 °C/min, the units of desorbed NH₃ increases compared to 10 °C/min, since the system receives more energy, which allows for more efficient desorption. These results agree with the available literature data [23]. Moreover, at higher heating rates, the desorption maximum shifts towards a higher temperature, especially for an LTA membrane. This behavior is often associated with restrictions on mass transfer within particles [23].

Investigation of the IR spectra of ammonia (Figure 6) adsorbed on the surface of the LTA and SOD zeolite membranes showed the presence of absorption bands in the range 3302–3307 and 3329–3331 cm^–1, which were assigned to asymmetric valence and deformation vibrations of the adsorbed ammonia, respectively, which forms a coordination bond with the proton centers of the surface [24]. Weaker absorption bands were attributed to the ammonium ion remaining on the surface of the zeolites. Molecular adsorption of ammonia occurs during the specific interaction of ammonia molecules with the surface of the zeolite with the formation of a bond. The absorption bands of molecular adsorbed ammonia were observed at about 3401–3423 cm^–1 on both types of membranes. The absorption band of ammonia molecules with a frequency of 3336 and 3342 cm^–1 is characterized by a coordination bond with a Lewis center. These bands were interpreted as asymmetric valence deformation vibrations of an ammonium ion [25,26,27,28,29,30,31]. The absence of absorption bands at 3650 and 3680 cm^–1 indicates that the corresponding hydroxyl groups are acid sites. Moreover, they confirm the presence of Lewis acid centers in the samples along with the Bronsted centers.

In this study, the acidity of each center was calculated, and the total acidity for both membranes was determined (

Table 1. The content of the acid sites on the surface of the SOD zeolite membrane, determined by the method of temperature-programmed desorption of NH₃.

Heating Rate β, °C/min	Temperature Maximum, Ti		Acidity Corresponding to Ni × 10−19, Units/m2		Total Acidity, Corresponding to ⅀N(i−j) × 10−19, Units/m2
10	T1	106	N1	0.00012 ± 0.000012	0.006 ± 0.0006
	T2	161	N2	0.00029 ± 0.000029
	T3	270	N3	0.0033 ± 0.00033
	T4	366	N4	0.00229 ± 0.000229
15	T1	108	N1	0.0001 ± 0.00001	0.0065 ± 0.00065
	T2	163	N2	0.0003 ± 0.00003
	T3	265	N3	0.0036 ± 0.00036
	T4	362	N4	0.0024 ± 0.00024
20	T1	107	N1	0.0001 ± 0.00001	0.0063 ± 0.00063
	T2	162	N2	0.0003 ± 0.00003
	T3	261	N3	0.0035 ± 0.00035
	T4	360	N4	0.0023 ± 0.00023
25	T1	105	N1	0.0001 ± 0.00001	0.0059 ± 0.00059
	T2	161	N2	0.0003 ± 0.00003
	T3	266	N3	0.0033 ± 0.00033
	T4	365	N4	0.0022 ± 0.00022
30	T1	108	N1	0.0001 ± 0.00001	0.0062 ± 0.00062
	T2	162	N2	0.0003 ± 0.00003
	T3	275	N3	0.0035 ± 0.00035
	T4	368	N4	0.0023 ± 0.00023

and

Table 2. The content of acid sites on the surface of the LTA zeolite membrane, determined by the method of temperature-programmed desorption of NH₃.

Heating Rate β, °C/min	Temperature Maximum, Ti		Acidity Corresponding to Ni × 10−19, Units/m2		Total Acidity, Corresponding to ⅀N(i−j) × 10−19, Units/m2
10	T1	82	N1	0.0120 ± 0.00120	0.08 ± 0.008
	T2	130	N2	0.0210 ± 0.00210
	T3	160	N3	0.0185 ± 0.00185
	T4	251	N4	0.0155 ± 0.00155
	T5	395	N5	0.013 ± 0.0013
15	T1	85	N1	0.0121 ± 0.00121	0.079 ± 0.0079
	T2	135	N2	0.0394 ± 0.00394
	T3	170	N3	0.0118 ± 0.00118
	T4	255	N4	0.0058 ± 0.00058
	T5	405	N5	0.0099 ± 0.00099
20	T1	95	N1	0.0036 ± 0.00036	0.0773 ± 0.00773
	T2	140	N2	0.0333 ± 0.00333
	T3	187	N3	0.0279 ± 0.00279
	T4	263	N4	0.0027 ± 0.00027
	T5	425	N5	0.0098 ± 0.00098
25	T1	101	N1	0.0058 ± 0.00058	0.0801 ± 0.00801
	T2	150	N2	0.0312 ± 0.00312
	T3	200	N3	0.0253 ± 0.00253
	T4	275	N4	0.0081 ± 0.00081
	T5	430	N5	0.0097 ± 0.00097
30	T1	105	N1	0.0067 ± 0.00067	0.083 ± 0.0083
	T2	150	N2	0.0363 ± 0.00363
	T3	201	N3	0.0235 ± 0.00235
	T4	280	N4	0.0076 ± 0.00076
	T5	435	N5	0.0089 ± 0.00089

). For the LTA zeolite membrane, its values are in the range of 0.08 × 10¹⁹ units/m² while for the sodalite membrane, its value is substantially lower at 0.006 × 10¹⁹ units/m². At the same time, the reliability of the results obtained is indicated by the fact that the total acidity does not depend on the heating rate of the samples; all the results are within the experimental error.

The TPD spectra of ammonia for LTA and SOD zeolites (Figure 7) have the same number of temperature maxima, nature, and number of peaks, but the total number of acid sites is 20–25% less compared to zeolite membranes of the same type. Various authors have studied the kinetics of ammonia desorption on zeolites [32].

An essential characteristic determining the sorption and catalytic properties of zeolites in acid-base reactions is the binding energy of ammonia with the sample surface [25,26].

Let us consider a more detailed calculation of the activation energy using the example of the most pronounced peak on the TPD curve for the LTA zeolite membrane (Figure 8) and the SOD membrane (Figure 9), the location of which depends on the removal rate in the temperature range of 350–500 °C. The temperature maxima for the LTA membrane, recorded at different heating rates of the spectra, are located as follows: T₁ (10 °C/min) = 395 °C, T₂ (15 °C/min) = 405 °C, T₃ (20 °C/min) = 425 °C, T₄ (25 °C/min) = 430 °C, and T₅ (30 °C/min) = 435 °C. Using the Kissinger coordinates ($\ln \frac{T_j^2}{\beta }$ oт $\frac{1}{T_j}$), we obtained a linear dependence, and the tangent of the slope angle of a given curve allowed us to calculate the activation energy [27,28]. For this case, Ea = 52.18 ± 2.42 kJ/mol. The calculation of the high-temperature peak for the sodalite membrane, the activation energy of which is 49.94 ± 3.99 kJ/mol, was carried out similarly.

The data resulting from the calculation of the activation energies for all TPD curves of the SOD and LTA zeolite membrane are summarized in

Table 3. Acidic properties of the SOD zeolite membrane obtained by NH₃ desorption.

SOD Membrane Heating Rate, °C/min	Temperature Maximum	Tmax, °C	Ea, kJ/mol
10	T1	106	12.74 ± 1.02
15		108
20		107
25		105
30		108
10	T2	161	18.04 ± 1.26
15		163
20		162
25		161
30		162
10	T3	270	32.3 ± 1.715
15		265
20		261
25		266
30		275
10	T4	366	69.94 ± 3.99
15		362
20		360
25		365
30		368

and

Table 4. Acidic properties of the LTA zeolite membrane obtained by NH₃ desorption.

LTA Membrane Heating Rate, °C/min	Temperature Maximum	Tmax, °C	Ea, kJ/mol
10	T1	82	20.24 ± 1.21
15		85
20		95
25		101
30		105
10	T2	130	28.31 ± 1.98
15		135
20		140
25		150
30		150
10	T3	160	35.72 ± 1.88
15		170
20		187
25		200
30		201
10	T4	251	74.26 ± 4.91
15		255
20		263
25		275
30		280
10	T5	395	78.18 ± 2.42
15		405
20		425
25		430
30		435

, respectively. It is worth noting that a directly proportional relationship between the strength of the acid sites and the activation energy values is observed [29]. For the LTA zeolite membrane in the temperature range of 50–200 °C, the activation energies are 20–35 kJ/mol (Table 4). This corresponds to the desorption of ammonia from α-cages. A further increase in the activation energy of the ammonia desorption to 70–80 kJ/mol is associated with the removal of ammonia from the less accessible β-cages of the LTA zeolite.

In contrast, the simpler and close-packed structure of the SOD membrane in the same temperature ranges shows lower values of the activation energy; namely, 10–20 kJ/mol for the processes of physical desorption of ammonia from the membrane surface, and 30–70 kJ/mol for the high-temperature regions (Table 3).

The data obtained are in good agreement with the fact that surface phenomena occur only when there is an excess of free energy in the boundary layer or in the presence of surface energy, which decreases proportionally to the surface area of the zeolite membrane [11,16,18]. Consequently, the qualitative characteristics of zeolite materials are directly proportional to changes in their surface area and structure [10,18]. Thus, the large specific surface area of the LTA membrane (150 m²/g) concerning the SOD membrane (115 m²/g), and the presence in the first case of the α and β cavities in the sample, providing access to OH groups, determines the increased content of Bronsted acid sites.

If we link the data on the content of the acid-base groups with the properties of adsorption/desorption of water vapor on these types of membranes, then we can say that the isotherms display hysteresis and belong to type IV, which indicates the presence of capillary condensation in the zeolite membranes (Figure 10). The maximum moisture capacity of the LTA zeolite membrane is about 60 wt%, which is 1.5 times that of the SOD membrane. These facts are explained by the presence of SOD in zeolite only as β-cavities while LTA zeolite also has α-cavities with a higher adsorption capacity. Consequently, it is not the developed surface of the sorbent that is critical for the adsorption process but the presence of regular cavities in the zeolite framework. A similar pattern for LTA and SOD zeolites is also observed regarding the processes of adsorption/desorption of water, which was confirmed in a previous study, where it was shown that the increased characteristics of LTA zeolite compared to sodalite are associated with the presence of alpha and beta cells in the structure of the first one [33].

4. Conclusions

In the course of this study, a comparative analysis of the acidic properties of the surface of LTA and SOD zeolite membranes was carried out. The membranes were obtained using ultrasonic treatment by applying a zeolite layer on Al₂O₃ supports and subsequent hydrothermal crystallization of the precursors obtained on the substrate surface. As a result of synthesis, defect-free well-crystallized membranes were formed. The surface area and the shape of the “assembly” of sodalite cages are the factors determining the presence of Bronsted acid sites on the surface of the samples and consequently, their reactivity is due to the surface area and the shape of the “assembly” of sodalite cages. The increased content of acid sites on an LTA membrane is determined by the presence of α and β cages in the structure, which facilitates access to additional OH groups in the sample and contributes to increased sorption characteristics when studying the properties of membranes for adsorption/desorption of water vapor.

The study results were confirmed by processing kinetic data based on the use of the Kissinger model. The apparent activation energy values for the desorption of ammonia from zeolite membranes with LTA and SOD were calculated. It was found that at temperatures below 250 °C, the degree of conversion, the activation energy is no more than 35 kJ/mol, which corresponds to the desorption of physically bound ammonia. An increase in the activation values up to 70 kJ/mol (for SOD) and up to 80 kJ/mol (for LTA) is associated with the desorption of chemically bound ammonia from the samples.