Natural gas is of significant importance as one of the highly consumed, efficient, and clean energy supply [1–3]. Accordingly, natural gas purification technologies have been developed extremely in the last decades. Among the hydrocarbon and non-hydrocarbon impurities present in natural gas, N2 is the most challenging one to remove [4]. Nevertheless, N2 removal is necessary since a high N2 content decreases the heating value. Therefore, a cost-effective technology is a must to obtain high-purity CH4 for meeting pipeline specifications; i. e., with a N2 concentration below 4% [3–5]. Currently, N2 is separated from CH4 by cryogenic distillation on industrial scale. It is not affordable as it is a high-priced and energy-consuming process, appropriate for large-scale processing of natural gas with high N2 content (i.e., > 10%) [6]. Research and industry sectors are trying to develop alternative and cost-effective processes that can be applied to both low-flow rate natural gas productions and streams containing below 10% N2. Among the proposed processes, e. g., solvent absorption and pressure swing adsorption (PSA), membrane technology can be an appropriate candidate due to its high energy efficiency, absence of phase change, easy maintenance, and limited number of process parameters (i.e., temperature and pressure) to be controlled during the process [7–10]. N2-selective and CH4-selective membranes are the two forms of N2 rejection membranes that are currently available. In the case of CH4-selective membranes, low-pressure CH4 will be obtained, which is not desirable since the need to obtain purified CH4 at elevated pressure (e.g. for transport through pipelines) involves recompressing costs. Therefore, research on appropriate N2-selective membrane is highly relevant [10]. Fabricating N2-selective membranes is demanding and challenging since N2 bears a strong resemblance to CH4 in physical and chemical properties [8]. In other words, difficulties in N2/CH4 separation by membrane or adsorption processes are connected to the proximity of N2 (3.64 Å) and CH4 (3.80 Å) kinetic diameters. Also, the higher polarizability of CH4 compared to N2 (26.0 ×10− 25 cm3 vs. 17.6×10− 25 cm3) causes a higher solubility in polymeric membrane materials and adsorption in porous solids, which conflicts with the need to obtain methane at elevated pressure [2, 12]. As a result, rubbery polymeric membranes are not appropriate for N2/CH4 separation since they are methane-selective. Also, acceptable N2 selectivity of glassy polymeric membranes has not been reported yet. For glassy polymers, the N2/CH4 selectivity is below 2.8 while the CH4/N2 selectivity is approximately 3 for rubbery membranes [12].
Although a wide range of industrial gas purification is dominated by polymeric membranes, over the past few years, there has been a significant interest in using molecular sieving inorganic membranes for N2/CH4 separation. For instance, carbon molecular sieve membranes (CMS) displayed N2/CH4 selectivities as high as 7.7; microporous aluminophosphate (AlPO) membranes indicated higher N2 permeances than CMS membranes but lower N2/CH4 selectivity of 4.7 [3]. Inorganic and particularly zeolitic membranes are more attractive for this separation due to uniform and small pore size and potential of size selectivity [3]. It has been noted that as compared to conventional materials, metal-organic frameworks (MOFs), covalent organic frameworks (COFs), and their ionic liquid-based composites (ILs) and polymers have been reported to have a noticeably highier potential for CH4/N2 separation [4, 13].
Small-pored zeolite membranes (e.g. with zeolites of AEI, DDR, and CHA topology) have great potential to achieve an acceptable selectivity for N2/CH4 separation [5]. AEI-type (AIPO-18) and CHA-type (SAPO-34 together with SSZ-13) zeolite membranes show good nitrogen permeances, which are located over the top bound of the Robeson 2008 plot (for polymeric membranes) [14]. Li et al. [11] reported the SSZ-13 membrane’s successful functioning (N2 permeance of 8.9 ×10− 7 mol.m− 2.s− 1.Pa− 1 and N2/CH4 permselectivity of 13.8 at 298 K and the feed pressure of 0.303 MPa). Substantial separation performances were also acquired by SAPO-34 membranes [15]. In another study, the N2/CH4 permselectivity of DD3R membrane was climbed above 20 at room temperature [16]. Among the inorganic membranes, the zeotype Engelhard Titanosilicate (ETS-4) membrane has gained significant attention in N2/CH4 separation due to its unique structural features. An ion-exchanged ETS-4 membrane, reported by Guen et al. [17], was N2-selective with N2/CH4 permselectivity of 3.8 at 308 K, and N2 permeance of about 1×10− 8 mol.m− 2.s− 1.Pa− 1. In another study, this group reported that the N2/CH4 permselectivity for titanosilicate membrane approached a value of 2.5 (N2 permeance of 2.5×10− 8 mol m− 2 s− 1 Pa− 1) [18].
ETS-4 is known as small-pored member of the titanosilicate zeolite-type materials, and this feature permits it to play a notable role as a size-selective adsorbent and membrane [19]. ETS-4 owns TiO6 octahedra and also corner-sharing SiO4 tetrahedra structure with 12 and 8MR pore openings. However, faults in the crystalline framework result in the blocking of 12MR pores (in the c-direction) in a way that the it has molecular sieving qualities akin to small-pore zeolites. In the b-direction, 8MR channels are the only penetrable paths for diffusing molecules [20, 21]. Therefore, this structure promotes the adsorption of smaller molecules including H2O, N2, NH3, H2S, and SO2 [22]. Another promising feature of ETS-4, which has been reported by Lin et al. [23], is the halogen anions replacement of OH groups inside the ETS-4 structure. Understandably, its pore size is engineering through the integration of halogen anions (i.e., Cl, I, and F). Such halogen anions can result in a partial roadblock of 8MR channels because TiO5 is projected into the 8MR pore openings. On the other hand, similar to zeolitic materials, extra-framework cations (i.e., Na+, and K+) of the ETS-4 give ionic features to its structure [23, 24], which strongly attract water molecules from ambient air. To remove those water molecules from the 8MR channels, the material has to be activated at elevated temperatures to free up the pore volume [26]. However, that activation temperature should be considered cautiously due to ETS-4’s poor thermal stability in its as-produced form and the risk of collapsing the structure near 200 ˚C [27].
Despite the potential of ETS-4 in N2/CH4 separation, only a few studies report on the performance of ETS-4 membranes in this process. Therefore, the objective of this study is to fabricate a highly N2-selective ETS-4 membrane on a porous α-alumina support tube by a secondary growth approach. Increasing the permselectivity of N2/CH4 is another purpose of current research. In this regard, ETS-4 seeding powder was synthesized by a novel method to obtain small-sized ETS-4 crystals. In fact, the use of small-sized seeding powder leads to high coverage of the substrate in the rub coating step. The membrane gel composition and synthesis conditions were modified to provide high inter-grown ETS-4 crystals to improve permselectivity. The produced powder and membranes were characterized by FESEM, XRD, and EDX techniques. The obtained membranes were thermally activated in the temperature range of 80–140 ºC before permeance measurements. The effect of such thermal activation on the ETS-4 membrane performance has not been reported in literature. Characteristics of gas permeation in the synthesized ETS-4 membranes were fully investigated for N2 and CH4 at 30, 50, and 70 ºC at a pressure difference up to 600 kPa.