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

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.


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 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: or in a linear form: 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: 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 NH 3 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 NH 3 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: where 6.022 × 10 23 is Avogadro's number; S(T maxi ) is the area under the corresponding maximum on the thermal desorption curve, mm 2 ; S sp is the specific surface of the support samples, m 2 /g; G is the weight, g; and ΣS (T maxi ) is the total area of peaks on the thermal desorption curve, mm 2 ; 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 H 2 O. 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.% NH 3 . After saturation, it was washed with He. N 2 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.
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-Al 2 O 3 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 Al 2 Si 2 O 7 /NaOH/γ-Al 2 O 3 = 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-Al 2 O 3 was additionally introduced into the initial reagent mixture. The optimal γ-Al 2 O 3 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). Al 2 O 3 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.

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 Figures 1 and 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 NH 3 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-NH 3 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 NH 3 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.  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 sorbed 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. It was found that the TPD spectra of ammonia desorbed from the surface of the sodalite membrane ( Figure 4) exhibit four desorption temperature maxima (T 1 -T 4 ). They can be conventionally ranked as weakly acidic (T 1 , T 2 ) and moderately acidic (T 3 ), and strong (T 4 ). 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 mem- 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. T 1 , T 2 , and T 3 also belong to the low-temperature region, which is characteristic of the desorption of physically adsorbed NH 3 molecules [22]. Ammonia is adsorbed on acid sites, which most likely have electron-withdrawing properties, based on the structural features of LTA zeolite (Na 12 Al 2 Si 12 O 48 ), 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 NH 3 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 (Tables 1 and 2). For the LTA zeolite membrane, its values are in the range of 0.08 × 10 19 units/m 2 while for the sodalite membrane, its value is substantially lower at 0.006 × 10 19 units/m 2 . 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].  In this study, the acidity of each center was calculated, and the total acidity for both membranes was determined (Tables 1 and 2). For the LTA zeolite membrane, its values are in the range of 0.08 × 10 19 units/m 2 while for the sodalite membrane, its value is substantially lower at 0.006 × 10 19 units/m 2 . 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: T 1 (10 • C/min) = 395 • C,  ), 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. stantially lower at 0.006 × 10 19 units/m 2 . 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]. Figure 7. Thermal desorption spectra of NH3 on SOD and LTA zeolites, recorded at a heating rate of 30 °C/min. Figure 7. Thermal desorption spectra of NH 3 on SOD and LTA zeolites, recorded at a heating rate of 30 • C/min. 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 ( pendence, 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 Tables 3 and 4, 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 m 2 /g) concerning the SOD membrane (115 m 2 /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. The data resulting from the calculation of the activation energies for all TPD curves of the SOD and LTA zeolite membrane are summarized in Tables 3 and 4, 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 m 2 /g) concerning the SOD membrane (115 m 2 /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]. 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].

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 Al 2 O 3 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.