A Study of the Influence of Synthesis Parameters on the Preparation of High Performance SSZ-13 Membranes

Abstract

This study investigated the effect of different synthesis parameters including pre- and post-hydrothermal treatment on the formation of a high-quality SSZ-13 membrane layer. The membranes were identified initially by the gas tightness test, then were characterized by single gas permeation measurements applying H₂, He, CO₂, N₂, CH₄, and SF₆ at room temperature. The results showed how each parameter affects the performance of the membrane, including structural defects in the formed selective layer, CO₂ permeance, and the ideal CO₂/CH₄ permselectivity. This work focused on optimizing these parameters. An ideal CO₂/CH₄ permselectivity of up to 122 with CO₂ permeance of ~3.72 × 10⁻⁶ [mol/(m²sPa)] and CO₂/CH₄ selectivity of 111 with CO₂ permeance of 8.5 × 10⁻⁷ [mol/(m²sPa)] in an equimolar mixture at room temperature and pressure drop of 0.15 MPa was achieved. This is one of the highest performances compared to other publications for SSZ-13 or all-Si membranes.

1. Introduction

One of the main factors in global warming and climate change issues is the presence of CO₂, which may be found in various sources such as natural gas [1], flue gas [2], biogas [3,4], and products of coal gasification [5]. To realize the importance of the issue, it must be considered that in the last five years more than 30 billion tons of CO₂ have been emitted, which has always been a growing amount. Therefore, CO₂ capture and separation has become one of the most important research topics in the present century [6,7,8,9]. There are currently a few technologies available for CO₂ separation including cryogenic distillation [10], amine absorption [11], amine scrubbing [12], etc., that are characterized by high maintenance costs and energy consumption, with energy requirements ranging from 4 to 6 MJ/kg CO₂ for absorption, 2 to 3 MJ/kg CO₂ for adsorption, and 6 to 10 MJ/kg CO₂ for cryogenic distillation. However, membrane technology, which does not require sorbent regeneration or desorption, is more economical compared to other technologies due to its advantages such as lower energy consumption (0.5 to 6 MJ/kg CO₂) and the ability to separate different mixtures [6,13]. As yet, different types of membranes have been used and studied for CO₂ separation such as metal organic frameworks (MOFs) [14] as well as polymeric [15], mixed-matrix [16], and inorganic membranes [17,18]. In recent years, polymeric membranes were pioneers in the CO₂ capture process, but their low stability in industrial conditions and also their low CO₂ permeance and CO₂/CH₄ selectivity in comparison with inorganic membranes have limited their use [19]. The separation industry requires membranes with high gas permeance, structural stability, and selectivity. Nowadays, zeolite membranes are attracting more attention because these membranes combine all the above-mentioned factors, especially desirable pressure tolerance and high chemical and thermal stability during the separation process. Small-pore zeolites, 8-member-rings such as CHA, DDR, and LTA have high performance for CO₂ capture due to the adsorption–diffusion mechanism [20,21,22]. The CHA zeolite framework includes various types such as all-silica CHA [23], SSZ-13 (high-silica CHA) [24], SAPO-34 (silicoaluminaphosphate) [25], and low-silica CHA (chabazite) [26,27]. Among these four types, SSZ-13 has better performance because, for instance, SAPO-34 has less mechanochemical stability in harsh media due to the presence of phosphate compounds [27]. In addition, due to its high CO₂ selective properties and hydrothermal stability, SSZ-13 membranes with a ~0.38 × 0.42 nm pore size are appropriate materials for usage in CO₂/CH₄ separations with high CO₂ permeance [21,28]. According to recent publications regarding high-Si CHA membranes for gas separation, the CO₂/CH₄ permselectivity is in the range of 20–300, but in most of the publications, the CO₂ permeance is less than 10⁻⁶ [mol/(m²sPa)] [19], because, usually, several cycles of hydrothermal synthesis are used to achieve a flawless membrane layer, which increases the transmission resistance [29]. In this work a detailed study of fabricating thin and uniform SSZ-13 zeolite membranes by controlling the many critical parameters in the synthesis procedure is presented. It is not only very important for CO₂/CH₄ separation in natural gas processing but also for SF₆ recovery (highly harmful greenhouse gas), that SSZ-13 membranes exhibit high separation selectivity for both applications. By analyzing various synthesis parameters in this study, it was possible to fabricate SSZ-13 zeolite membranes with high structural stability in a reproducible method, making them appropriate for upscaling in gas separation applications.

2. Materials and Methods

To synthesize crystals and membranes in this work, all chemicals were used without further treatment, including Ludox^® AS-40 (40 wt%, colloidal silica, Sigma-Aldrich, Steinheim, Germany), tetraethylorthosilicate (TEOS, 98 wt%, Merck, Darmstadt, Germany) and fumed silica (CAB-O-SIL^® M5, Cabot, Frankfurt am Main, Germany) as sources of Si, aluminum hydroxide (Al(OH)₃, extra pure, ACROS organics, Geel, Belgium), aluminum isopropoxide (IPA, 99 wt%, Sigma-Aldrich, Steinheim, Germany) and sodium aluminate (NaAlO₂, ~45 wt% Na₂O and ~55 wt% Al₂O₃, Sigma-Aldrich, Steinheim, Germany) as Al sources, and sodium hydroxide (NaOH, reagent grade ≥ 98%, pellets, anhydrous, Merck, Darmstadt, Germany), as a source of alkali were used. As organic structure directing agent (OSDA), N,N,N-trimethyl-1-adamant ammonium hydroxide (Sachem Inc., TMAdaOH, 25 wt% aqueous in H₂O, Austin, TX, USA) was used. An ELGA PURELAB Pulse purification system was used to purify deionized water (1 s/cm) for all experiments. Lastly, high-purity gases were used to measure gas permeation—H₂ (99.9995%), He (99.9995%), N₂ (99.9995%), CO₂ (99.998%), CH₄ (99.5%), and SF₆ (99.9%).

2.1. Preparation of Seed Crystal

In reference to [30], SSZ-13 crystals with a molar ratio of 1 SiO₂:0.2 NaOH:0.05 Al:0.2 TMAdaOH:44 H₂O were prepared. Some minor modifications were made, including reducing the TMAdaOH/Si ratio from 0.4 to 0.2. The chemical components were sequentially mixed, and the homogeneous mixture was subjected to an overnight aging process. Afterward, the resulting mixture was transferred into a Teflon-lined autoclave and treated at 160 °C for a duration of 100 h. The products were collected through centrifugation and then washed with DI-water until approximately pH 7 was reached. The obtained CHA crystals were utilized as seeds for the synthesis of SSZ-13 membranes.

2.2. Seeding Procedure

As support for the membrane layer, porous tubular α-alumina (105 mm length, 10 mm outer diameter, 7 mm inner diameter and membrane surface area of ~17 cm², with 100, 200, 800 nm and 3 µm average pore size in the final supporting layer) was used. Tubes were sealed with glaze and then dip-coated with seed crystals (0.05–2 wt% concentration). The dip coating was repeated twice in different directions to form an even seed layer. After the seeding procedure and drying at 100 °C, the outer surface of the tube was covered with PTFE to avoid zeolite attaching and growth. Finally, membranes were synthesized on the inside surface of the tubes via secondary growth.

2.3. Hydrothermal Synthesis of SSZ-13 Zeolite Membrane

The molar composition of the synthesis gel was 1 SiO₂:0.W TMAdaOH:0.X NaOH:0.Y Al(OH)₃:Z H₂O and the values of W, X, Y, Z were varied from 0.1 ≤ W ≤ 0.5, 0.1 ≤ X ≤ 0.2, 0.1 ≤ Y ≤ 0.4 and 20 ≤ Z ≤ 80. The gel was prepared by stirring NaOH, Al(OH)₃, and H₂O for 2 h. The template (TMAdaOH) was added and stirred for another 1 h and finally, the colloidal silica sol was added to the mixture and the solution was aged for 2–24 h at room temperature. The seed-coated substrates were placed vertically in the autoclave. Then, a synthesis temperature of 140–180 °C, as well as synthesis times of 24–72 h, were used. After unwrapping the Teflon tape, the as-synthesized SSZ-13 membrane was rinsed and soaked in DI water for 24 h and dried at 100 °C for 12 h.

2.4. Gas-Tightness Measurement

For a high-quality SSZ-13 membrane, achieving gas-tightness before calcination (to remove OSDA) is crucial. Pre-calcination, N₂ permeance can be attributed to diffusion through defects and non-zeolitic pores since the zeolitic pores in the SSZ-13 framework are filled with the organic template. This N₂ permeance serves as a predictor of membrane quality. If the N₂ permeance approaches zero prior to calcination, it indicates suitability for further steps (template removal). Conversely, a high N₂ permeance suggests the presence of defects or cracks on the membrane surface.

2.5. Template Removal of SSZ-13 Membranes

The membranes were calcined to eliminate the OSDA in air at temperatures ranging from 400–600 °C for 6 h, with heating and cooling rates of 0.5 K/min.

2.6. Single and Mixed Gas Permeation Measurement

Membrane characterization utilized a homemade single gas permeation setup. Various gases, including He, H₂, CO₂, N₂, CH₄, and SF₆, were measured in dead-end mode using the pressure rising method. The kinetic molecular diameter of gases can be observed in Figure S1 [30,31,32]. In this study, single gas permeation measurements were performed three times for each gas, with repeated cycles of separate vacuum, gas flow, and average values being reported. Details of the single gas permeation measurement set up are described in Figure S2. CO₂/CH₄ mixtures were separated using mass flow controllers that controlled flow rates. An overpressure of 0.005 MPa was maintained on the permeate side at a pressure of 0.105 MPa (absolute) and no sweep gas was introduced. Permeate flow was measured using a soap bubble counter, while feed gas flow was controlled by a mass flow controller. A gas chromatographic method was used to analyze feed, retentate, and permeate compositions. Selectivity, a unitless factor, reflects the ability of the membrane to separate components A and B in a gas mixture. If the partial pressure of the gas mixture (comprising components A and B) introduced on the feed side (high-concentration) of the membrane is denoted as X_A and X_B, and the partial pressure on the permeate side (low-concentration side) is Y_A and Y_B, the separation factor α is defined as follows [9]:

$$\alpha ={\displaystyle \frac{Y_A/Y_B}{X_A/X_B}}={\displaystyle \frac{Y_A{X}_B}{Y_B{X}_A}}$$

(1)

The permeance P_A of component A in a gas mixture is defined as the amount n_A of gas A that permeates through a specific membrane area A over a specific time t, driven by the partial pressure difference p_A of gas A across the membrane. The permeance is given by the following:

$$Permeance P_A={\displaystyle \frac{n_A}{At{p}_A}}$$

(2)

and is expressed as [mol/m²sPa] in SI units [17].

3. Results and Discussion

3.1. Impact of Pore Size in Intermediate Layers

Asymmetric membranes consist of various layers, including a membrane layer, intermediate layers, and a support that enhances the mechanical stability of the upper layers. More details about the substrate can be found elsewhere [32]. The pore size of the intermediate layer significantly affects the diffusion process of the membrane, as the thin top layers must be uniformly formed on it without defects (Figure 1a). Regarding this parameter, alumina supports with pore sizes of 100 nm, 200 nm, 800 nm, and 3 μm were used as intermediate layers. These membranes were fabricated under identical conditions, and based on the outcomes in Figure 1b, it is noticeable that the intermediate layers with size of 800 nm and 3 μm display high N₂ permeance (1.31 × 10⁻⁵ and 7.28 × 10⁻⁵ [mol/(m²sPa)], individually), from which it can be inferred that the membrane layer exhibits incomplete coverage across the surface due to the presence of intercrystalline gaps. Thus, these gaps provide insufficient nucleation sites for the formation of a compact and homogenous SSZ-13 layer and permit N₂ molecules to penetrate through flaws and non-zeolitic pathways. Furthermore, an increase in the intermediate layer pore size leads to a thicker membrane due to enhanced capillary filtration. This increase in layer thickness increases the possibility of structural defects in the membrane even before the calcination process [20,21]. Nevertheless, the N₂ permeance for intermediate layers of 100 and 200 nm decreases to 4.96 × 10⁻⁷ and 1.24 × 10⁻⁹ [mol/(m²sPa)], respectively. The support with a pore size of 100 nm has defects (non-zeolitic pores, because seed adhesion and penetration were not sufficient to penetrate the surface) that affect membrane efficiency. Based on the outcomes, intermediate layers with pore sizes of 200 nm formed a tight layer and were selected for further investigations.

3.2. Impact of Seeding Concentration

The seed suspension was prepared by dispersing SSZ-13 crystals in ethanol (EtOH) at different concentrations and to prevent agglomeration subjected to ultrasonic treatment for 30 min. Generally, as the concentration of the seed suspension increases, the thickness of the formed membrane layer also increases (due to the nucleation site) [33]. The optimal seed concentration ensures the formation of a homogeneous and dense zeolite layer by creating sufficient nucleation sites throughout the substrate. However, high nucleation sites can lead to agglomeration, resulting in a thicker and potentially less uniform layer. Conversely, a lower concentration of seed crystals may result in incomplete coverage and create a non-uniform layer that cannot adequately cover the entire surface. It is necessary for the membranes to have minimum thickness, because by maintaining selectivity the permeance increases and the possibility of defects or cracks in later steps, e.g., the calcination operation, is reduced. Based on Table S1, membranes with a seed suspension concentration less than 0.25 wt% and above 1 wt% showed N₂ permeances higher than 7 × 10⁻⁹ [mol/(m²sPa)] during the tightness test. After calcination, single gas permeation was measured on the membranes with seed concentrations of 0.25, 0.5, 0.75, and 1 wt%, as shown in Figure 2a. It can be seen with the increase in seed concentration from 0.25 to 1 wt%, the gas permeance reduced, due to the formation of a more uniform and denser layer and, as a result, permselectivity of H₂/SF₆, CO₂/N₂, and CO₂/CH₄ increased to 95, 14, and 57, respectively.

3.3. Impact of Sonication and Suspension Liquid on Deagglomeration of Seeds

To explore the influence of sonication treatment on the membrane’s performance, three suspensions with a concentration of 1 wt % were prepared in different solvents including EtOH, DI water, and a mixture of EtOH and DI water (50–50 wt%), then sonicated for 30 min. Detailed solvent specifications can be found in Table S2 [34]. The results showed that sonication could efficiently diminish gaps and defects in the formed layer due to deagglomeration and good dispersion, therefore improving penetration and adhesion to the support surface. While CO₂ permeance remained relatively constant, based on the results in Figure 2b, SF₆ permeance exhibited variability influenced by the suspension liquid and sonication time. Furthermore, membranes prepared using ethanol-based seed suspensions and subjected to 30 min of sonication demonstrated superior performance. The lower surface tension of EtOH facilitated uniform seed crystal dispersion, resulting in a homogeneous layer on the support surface, as represented in Figure S3. As indicated by the results, the permselectivities for the gas pairs CO₂/CH₄, CO₂/N₂, and H₂/SF₆ were calculated as 80, 20, and 90.

3.4. Impact of Sources of Si and Al

Si and Al sources critically impact the crystallization and nucleation of zeolites due to their dissolution rate (reactivity and solubility) during synthesis and ageing processes. SSZ-13 membranes can be synthesized utilizing diverse sources of Si and Al, e.g., colloidal silica Ludox^® AS-40 [35], monomeric tetraethylorthosilicate (TEOS) [36], fumed silica Cab-O-Sil^® M5 [37], aluminum isopropoxide [38], sodium aluminate [29], and aluminum hydroxide [39], respectively. During this study, three different sources of Si and Al were used as shown in Table S3. The gas tightness test revealed that only membranes prepared using aluminum isopropoxide in the gel precursor exhibited significant N₂ diffusion (1.2 × 10⁻⁷ [mol/(m²sPa)]), indicating a high defect density. Figure 2c demonstrated the superior CO₂/CH₄ separation performance of membranes synthesized using Ludox^® AS-40 and Al(OH)₃, since Al(OH)₃ is less soluble compared to precursors like Al-isopropoxide or sodium aluminate and leads to slower and more controlled release of Al ions during synthesis that can favor the formation of a more ordered and well-defined zeolite framework. That is, Al sources with different surface areas can lead to dissimilar crystallization rates (kinetics of dissolution during zeolite formation and growth), crystal dimensions, and microstructural characteristics. Moreover, Al(OH)₃ is more stable and has a lower reaction temperature. In addition, membranes synthesized from silica suspensions revealed nearly the same values in both CO₂ and SF₆ permeance. Nevertheless, when fumed silica was used, CO₂ permeance was remarkably decreased. Mainly, Si sources with lower surface area and solubility like fumed silica promote the growth of large crystals (and may lead to the formation of unreacted Si particles), while high surface area Si sources (e.g., colloidal silica) induce the growth of small crystals and can be effectively dispersed within gel precursors [40].

3.5. Impact of Si/Al Ratio

The mechanochemical characteristics, morphology, framework charge, stability, and hydrophobicity of the formed zeolite membrane layer are impacted by the Si/Al ratio in the gel precursor. Nevertheless, there is no measurable correlation between the Si/Al ratio of the initial reaction system and the product after the hydrothermal reaction [41]. According to Table S4, to achieve the optimal Si/Al ratio, five different ratios were investigated. At Si/Al = ∞, N₂ permeance is significantly high, which is representative of non-zeolitic pores, because a fully efficient all-Si CHA membrane is typically either synthesized in fluoride media [42] or the synthesis time is more than 48 h [43]. According to Figure 2d, as the Si/Al ratio in the precursor gel increases up to 20, the SF₆ permeance reduced (e.g., Si/Al ratio of 40, which showed the lowest concentration of defects and non-zeolite pores), which is similar to other publications [41]. Consequently, a Si/Al ratio of 20 exhibited maximum CO₂/CH₄, CO₂/N₂, and H₂/SF₆ permselectivity of 80, 19, and 85, compared to other ratios, resulting in minimum CH₄ permeance. This effect is due to the increased Si/Al ratio, which decreases both intercrystalline defects and the polarity of the SSZ-13 framework. Also, the Si/Al ratio of 10 results in the highest SF₆ permeance and lowest CO₂/CH₄ permselectivity.

3.6. Impact of Alkalinity (OH/Si)

Zeolite frameworks (e.g., aluminosilicate) typically crystallize under high alkaline conditions, characterized by a pH range of 9–13 [44]. The negative charge in the zeolite framework can be counterbalanced by cations such as Na⁺, K⁺, and Cs⁺ [45]. An elevated OH⁻/Si ratio promotes the solubility of Si and Al sources within the precursor gel, concurrently reducing the polymerization degrees of silicate anions and accelerating the polymerization of aluminate anions. In particular, increasing alkalinity accelerates zeolite crystallization by shortening the induction and nucleation periods (since the reaction between Al and Si ions at higher pH is accelerated by releasing OH⁻). Additionally, it affects the size and morphology of the crystal and expands the ionic interaction between tetrahedra and OSDA [46]. To study this parameter, five membranes were prepared with different OH⁻/Si ratios from 0.1 to 0.5. As outlined in Table S5, the membrane with the highest alkalinity (OH⁻/Si ratio of 0.5) displayed a pronounced presence of non-zeolitic pathways prior to calcination. Gas permeation measurement revealed that a reduction in the alkalinity of the gel precursor led to a slight decrease in CO₂ permeance but a significant drop in SF₆ permeance (from 0.2 to 0.05 [mol/(m²sPa)]), resulting in an increased ideal CO₂/CH₄ permselectivity from 17 to 85 (Figure 2e). In addition, the obtained permselectivity for CO₂/N₂ and H₂/SF₆ increased to 14 and 60, respectively.

3.7. Impact of H₂O/Si

The H₂O/Si ratio influences the ultimate characteristics of the membrane layer through the crystallization process and chemical reactivity (e.g., via changing the crystallization medium and concentration of reactants, respectively) [47]. In this respect, five membranes with H₂O/Si ratios of 20, 40, 60, 80, and 100 were synthesized under similar conditions. N₂ permeance measurement revealed high non-zeolitic pathways in the membrane with an H₂O/Si ratio of 100. According to Figure 2f by increasing the H₂O/Si ratio, the ideal CO₂/CH₄, CO₂/N₂, and H₂/SF₆ permselectivity increases from 0.6, 0.75, and 8 (for H₂O/Si = 20) to 94, 10, and 115 (for H₂O/Si = 80), respectively. In addition, the SF₆ permeance decreases from 1.98 × 10⁻⁷ to 8.68 × 10⁻⁹ [mol/(m²sPa)]. The noteworthy point is that for membranes with H₂O/Si ratios of 40, 60, and 80 the CO₂ permeance is higher than other gas molecules, but for a H₂O/Si ratio of 20, the CO₂ permeance is lower than that of He, H₂, N₂, and even CH₄. This phenomenon can be ascribed to the presence of large size defects and cracks. The CO₂/CH₄ permselectivity of ~0.6 confirms that based on Graham’s law ${\alpha }_{{\displaystyle \frac{CO2}{CH4}}}=\sqrt{{\displaystyle \frac{M_{CH4}}{M_{CO2}}}}$ [7], the membrane with H₂O/Si ratio of 20 presents a Knudsen diffusion mechanism, which is possibly due to the high concentration of the gel precursor leading to the creation of multiple nucleation sites, and as a result the formation of a thick layer and the emergence of structural defects during calcination.

3.8. Impact of TMAdaOH/Si

By optimizing the TMAdaOH/Si ratio, the crystallinity (ordered framework), membrane layer thickness, and intergrowth crystals can be improved to achieve a membrane with higher performance and lower defect density [21]. To investigate, different TMAdaOH/Si ratios were used as listed in Table S6. After calcination, membranes with TMAdaOH/Si ratios of 0.1, 0.15, 0.2, and 0.5 were selected for gas permeation measurement. As displayed in Figure 2g, with increasing TMAdaOH ratio, the permeance of gas molecules increased (especially for large molecules e.g., SF₆ and CH₄), and the CO₂/CH₄ permselectivity reduced from 104 to 4. In general, the crystal size increases with increasing TMAdaOH content. However, a high TMAdaOH/Si ratio in the gel composition is associated with increased defect formation during the template removal step. Furthermore, the increase in alkalinity caused by OSDA can lead to the dissolution of zeolite layers and afterwards reduce crystallite intergrowth [6]. The literature on SSZ-13 membranes commonly reports high TMAdOH/Si ratios, ranging from 0.19 [29], 0.2 [20,35,48], 0.5 [38,39], and finally 0.6 [21,49]. In contrast, this study achieved high CO₂/CH₄ permselectivity with reduced SF₆ permeance using a significantly lower TMAdOH/Si ratio of ~0.1 (minimizing OSDA consumption and associated costs while mitigating environmental impact).

3.9. Impact of Ageing Time

The aging process enhances the kinetics of zeolite nucleation and crystal growth (via increasing the dissolution of aluminosilicate species). According to Table S7, the synthesized membrane without ageing showed high N₂ permeance. According to Figure 2g, as the aging time increased from 2 to 24 h, the ideal CO₂/CH₄, CO₂/N₂ and H₂/SF₆ permselectivity also improved from 15 to 110, 11 to 14, and 20 to 56, individually. Nevertheless, maximum CO₂ and minimum SF₆ permeances were correlated to the samples aged 16 and 8 h (~2.97 × 10⁻⁶ [mol/(m²sPa)] for CO₂ and ~8.06 × 10⁻⁹ [mol/(m²sPa)] for SF₆, separately). The obtained results display that with the increase of ageing time, because of higher nucleation rate as well as shorter induction and crystallization period, the number of crystals on the surface of the support increased, while their size decreased. Moreover, the formation rate of aluminosilicate hydrogel is enhanced by the ageing process and hence, it is probable that by increasing the ageing time from 8 and 16 to 24 h, a denser membrane layer was formed, and this led to a decrease in CO₂ permeance.

3.10. Impact of Filling Factor of the Autoclave

The degree of fill can influence multiple features of the synthesis process, such as zeolite crystal nucleation and growth, crystallization rate (i.e., uniformity), crystal size (i.e., thickness), and morphology. Furthermore, the pressure inside the autoclave can impact the reaction rate and equilibrium [50,51]. To assess the influence of internal pressure on membrane performance, four fill factors (60%, 70%, 80%, and 90%) were employed. The results are depicted in Figure S4. According to Figure 3a, gas permeation measurement showed that as the filling factor increased from 60% to 90%, the ideal CO₂/CH₄ permselectivity (from 110 to 30), the ideal H₂/SF₆ permselectivity (from 125 to 31), and CO₂ permeance (from 2.85 × 10⁻⁶ to 1.73 × 10⁻⁶ [mol/(m²sPa)]) decreased and the permeance of SF₆ molecules increased (from 7.44 × 10⁻⁹ to 1.73 × 10⁻⁸ [m³/(m²sPa)]). The observed behavior can be attributed to increased pressure within the autoclave due to suspension expansion during the hydrothermal process, and this elevated pressure results in a more compact and stressed SSZ-13 membrane layer. Consequently, an elevated fill factor correlates with an increased incidence of non-zeolitic pathways. Moreover, heightened reactant concentration due to reduced free volume inside the autoclave can accelerate crystal growth, potentially leading to larger crystals but also increasing the risk of agglomeration or second phase formation.

3.11. Impact of Synthesis Time

To explore the influence of synthesis time, membranes were prepared over 24, 36, 48, 60, and 72 h. Gas-tightness testing (Table S8) revealed incomplete membrane formation across the substrate surface after 24 and 36 h (N₂ permeance of 6.78 and 0.25 × 10⁻⁷ [mol/(m²sPa)], respectively). In other words, inadequate crystal intergrowth, resulting in intercrystalline gaps, was observed for shorter synthesis times. Additionally, the presence of unreacted Si within the framework composition due to insufficient reaction time can negatively impact membrane performance. Nevertheless, by increasing the duration of synthesis to 48 h, the SSZ-13 layer was formed on the substrate surface (N₂ permeance of 0 [mol/(m²sPa)]). Based on Figure 3b, high ideal CO₂/CH₄ permselectivity at a synthesis time of 48 h is attributed to the formation of a well-interconnected membrane structure. Extending the synthesis time to 72 h, resulted in decreased ideal CO₂/CH₄ and H₂/SF₆ permselectivity from 103 to 39 and from 125 to 76, respectively, due to increased membrane thickness, hindering CO₂ permeance, and the potential formation of undesired phases [47].

3.12. Impact of Synthesis Temperature

The influence of synthesis temperatures ranging from 140 to 180 °C was evaluated, while tightness results revealed that lower temperatures (140 and 150 °C) were insufficient for the uniform formation of a selective layer, as evidenced by N₂ permeance through the selective layer of 19.01 and 0.03 × 10⁻⁷ [mol/(m²sPa)], separately (Table S9). Elevated hydrothermal temperatures led to a decrease in N₂ permeance, signifying the enhanced quality of the SSZ-13 layer. Figure 3c illustrates a decline in CO₂/CH₄ and H₂/SF₆ permselectivity from 122 to 45 and from 144 to 51, separately, as well as a decline in CO₂ permeance from 30 to 12 × 10⁻⁷ [mol/(m²sPa)]) with increasing synthesis temperature from 160 °C to 180 °C. Furthermore, SF₆ permeance increases with rising temperature and is similar to extended synthesis time, leading to thicker membranes, often accompanied by increased defect formation and potential second phase incorporation. This can be described such that at higher synthesis temperatures, the crystal growth rate will be enhanced (thicker membrane layer) and these intergrown structures will produce more flaws during calcination. In addition, elevated temperatures may cause second phases or impurities, resulting in low separation performance.

3.13. Impact of Calcination Temperature

Calcination treatment plays a critical role in the performance of SSZ-13 membranes since the thermal expansion coefficient of the alumina substrate is +8.7 × 10⁻⁶ K (for temperatures ranging from 25 to 800 °C), while the thermal expansion coefficient for the SSZ-13 membrane layer is −28.5 × 10⁻⁶ K (for temperatures from 25 to 600 °C, the negative value is due to the removal of OSDA from the framework structure), and this variance may cause defects or cracks during heat treatments, e.g., template removal [49]. To eliminate OSDA from the membrane composition, a calcination process was conducted at temperatures ranging from 400 to 600 °C for 6 h with a heating and cooling rate of 0.5 K/min (Table S10). According to the results, elevated calcination temperatures led to an increase in thermal stress and defects in the membrane layer, as evidenced by high SF₆ permeance and diminished ideal CO₂/CH₄ permselectivity. Additionally, based on Figure 3d, elevated calcination temperatures over 550 °C, resulted in a loss of preferential CO₂ adsorption properties, as evidenced by increased He and H₂ permeance compared to CO₂. Based on the outcomes regarding this section, the permselectivity of CO₂/CH₄, CO₂/N₂, and H₂/SF₆ were obtained as 121, 20, and 150, respectively.

3.14. Mixed Gas Permeation Measurement

Equimolar CO₂–CH₄ gas mixture measurement was carried out at room temperature for two SSZ-13 membranes at two feed pressures of 0.255 and 0.605 MPa. According to Figure 4, the results showed that by increasing the feed pressure from 0.255 to 0.605 MPa, the CO₂ permeance for the membrane M1 decreased from 8.5 × 10⁻⁷ to 3.5 × 10⁻⁷ [mol/(m²sPa)] and for the membrane M2 from 3.5 × 10⁻⁷ to 1.2 × 10⁻⁷ [mol/(m²sPa)] (the membranes labeled as M1 and M2 correspond to the membranes that exhibited the highest CO₂/CH₄ permselectivity in single gas permeation measurements, with values of 120 and 110, respectively). CH₄ permeance increased from 7.6 × 10⁻⁹ to 1.3 × 10⁻⁸ [mol/(m²sPa)] for the M1 and from 2.6 × 10⁻⁹ to 3 × 10⁻⁹ [mol/(m²sPa)] for the M2. Therefore, the CO₂/CH₄ selectivity decreased from 111 and 116 for two membranes at pressure 0.255 MPa to 27 and 46, respectively, at 0.605 MPa. These results showed that the permeance of CO₂ and CO₂/CH₄ selectivity decreased in the mixture compared to the single gas measurement which can be attributed to competitive adsorption (CO₂ molecules are hindered by CH₄ molecules) as well as molecular interactions in a binary mixture. In addition, at higher feed pressures, performance of membranes decreased due to higher competition for adsorption sites, pore blockage, and more diffusion of CH₄ molecules through defects (greater contribution of non-zeolitic pathways).

3.15. Comparison with Related Work

Table 1. Comparison of single and mixed gas permeation measurement data and thickness for SSZ-13 membranes in recently published studies with this work.

Support Type and Pore Size	Synthesis Condition			Pressure Drop	Membrane Characteristics
					Thickness	Single Gas		Mixed Gas
	Gel Precursor Molar Composition	Time	Temperature			CO2 Permeance	Ideal CO2/CH4 Selectivity	CO2 Permeance	CO2/CH4 Selectivity	Ref.
[nm]	SiO2:Al2O3:Na2O:TMAdaOH:H2O	[h]	[K]	[MPa]	[μm]	10−7 × [mol/(m2sPa)]		10−7 × [mol/(m2sPa)]
Al2O3, 300	1.05:005:0.1:0.2:44	144	433	0.6	4 to 6	3	20	2.8	36	[52]
Mullite, 130	1:0.005:0.2:0.1:80	-	-	0.4	-	3.2	34	-	-	[53]
Al2O3, 200	1:0.05:0.1:0.5:80	48	433	0.2	6	8.1	75	5.6	57	[39]
Al2O3, 200	1:0.1:0.1:0.6:44	240	373	0.14	0.67	43	123	39	162	[54]
Al2O3, 200	1:0.01:0.1:0.6:44	48	433	0.14	1.7	-	-	14	116	[21]
Al2O3, 100	1:0.005:0.1:0.4:44	24	453	0.2	1	1.1	91	13.2	125	[55]
Al2O3	1:0.025:0.2:0.2:80	72	433	0.2	3	-	-	2	22	[56]
Al2O3	1:0.005:1:1:80	48	433	0.2	3 to 4	-	-	2.9	56	[57]
Al2O3, 200	1:0.01:0.1:0.6:44	48	453	0.14	3.27	-	-	3.8	135	[58]
Al2O3, 300	1:0.025:0.1:0.07:100	20	433	0.2	3	5.1	-	5.3	240	[59]
Al2O3, 100	1:0.025:0.1:0.4:44	96	433	0.3	6	15	45	-	-	[60]
Al2O3, 200	1:0.01:0.1:0.6:44	2	473	0.2	1	11	138	5.5	112	[49]
Al2O3, 150	1.05:0.004:0.2:0.2:44	72	433	0.2	5.3	4	130	2.5	119	[20]
Al2O3, 80	1:0.005:0.09:0.19:42.75	18	433	0.2	2	3.56	31	3.4	30	[29]
Al2O3	1:0.025:0.2:0.1:80	48	433	0.2	3	6.5	37	0.87	100	[28]
Al2O3, 200	1:0.05:0.05:0.2:80	72	453	0.2	3.5	0.23	46	9	34	[35]
Al2O3, 200	1:0.005:0.1:0.2:40	96	453	0.2	2	10.1	108	12.1	183	[61]
Al2O3, 200	1:0.025:0.1:0.1:80	48	433	0.15	2	37.2	122	8.5	111	This work

compares the synthesis parameters, thickness, and CO₂/CH₄ separation performance of this study to other SSZ-13 membranes recently published. Although the membranes were fabricated with only one hydrothermal synthesis cycle, with high stability (the results of the stability test in the presence of water vapor can be found in the Supplementary Material in Figure S5), and with high reproducibility, they showed high CO₂ permeance and CO₂/CH₄ selectivity in comparison with other publications.

4. Conclusions

Several parameters influence the formation of SSZ-13 membranes and any change in these parameters would lead to a different form and, ultimately, a different membrane performance. In this work, all the effective factors in membrane synthesis were examined and the following was found:

• The membrane was synthesized on an alumina support with a pore size of 200 nm
• A seeding procedure was employed with 1 wt% concentration of the SSZ-13 crystals in EtOH followed by 30 min of sonication
• The gel composition had the ratio of 1 SiO₂:0.1 TMAdaOH:0.2 NaOH:0.05 Al(OH)₃:80 H₂O after 24 h of ageing
• Crystallizing at 160 °C for 48 h and calcining at 450 °C yielded the highest CO₂/CH₄ separation performance.

Optimizing these parameters results in an ideal CO₂/CH₄, CO₂/N₂, H₂/SF₆ permselectivity of 122, 20, and 150, respectively, as well as CO₂ permeance of ~3.72 × 10⁻⁶ [mol/(m²sPa)] and CO₂/CH₄ selectivity of 111, with CO₂ permeance of 8.5 × 10⁻⁷ [mol/(m²sPa)] in an equimolar mixture at room temperature and pressure drop of 0.15 MPa. Finally, it should be mentioned that several factors have a greater impact than others when it comes to membrane formation, such as temperature and synthesis time, seeding procedure, and molar composition, but the changes in one parameter can also affect the others.