Advances in membrane technology for nitrogen-methane separation with focus on design performance and future trends

One of the critical components of the world’s energy supply is natural gas, which happens to be the cleanest form of fossil fuel. It is mostly composed of methane (CH₄) at an average composition of 70–98% [1, 2], which, when combusted, produces carbon dioxide and water vapour. In contrast, other fossil fuels, such as coal and oil, which are complex molecules with a higher carbon-to-hydrogen ratio, nitrogen, and sulfur, produce a higher percentage of carbon dioxide and toxic gases such as sulfur dioxide, nitrogen oxides, carbon monoxide, and so on. Natural gas has a considerably smaller global market than oil since transporting it is difficult and expensive [3]. With an average annual growth of 1.8% since 2007, the world’s consumption of natural gas, which peaked at 100 trillion cubic feet in 2004, has been increasing more quickly than that of any other fossil fuel [4]. Given their huge reserves and global distribution, unconventional natural gas resources may become the most important basis for the long-term expansion of the natural gas industry. Conventional natural gas resources will be insufficient to supply demand due to the global transition from carbon-intensive fossil fuels to greener energy sources. Meanwhile, as supplies are depleted, conventional natural gas extraction will become increasingly challenging [5, 6]. Because of its large reserves and widespread distribution, unconventional natural gas resources may emerge as the main energy source for the natural gas industry’s long-term growth. However, the separation of a considerable number of impurities in unconventional natural gas has become a significant technical challenge for large-scale usage, particularly the high-efficiency separation of nitrogen contaminants [7]. To meet the long-term energy demand, these untampered reservoirs will have to be used; likewise, the significant development in energy-efficient separation technologies of contaminants needs to be observed.

The contamination of nitrogen above a certain level makes many natural gas reservoirs/sources unusable simply because they do not meet the pipeline specification of < 4% nitrogen content [3]. The presence of nitrogen in natural gas significantly reduces its heating value, reducing its overall energy efficiency and economic viability as a fuel source [8]. This highlights the critical need for the development and optimization of advanced separation techniques capable of effectively isolating nitrogen from methane, thereby improving fuel quality and ensuring its suitability for various applications. The removal of N₂ greatly limits the usage of high-quality natural gases, as the separation of CH₄ and N₂ is difficult due to their very close kinetic diameters and thermodynamic properties [9]. Among the various industrial gas separations, the separation of CH₄ from N₂ remains particularly challenging due to their similar kinetic diameters (3.80 Å vs. 3.64 Å) and non-polar nature, making size-based and solubility-based discrimination difficult [9]. This separation is critical in the treatment of nitrogen-contaminated natural gas (up to 50% N₂ in some reservoirs) and biogas purification, where CH₄ enrichment is required. Several approaches have been developed, each giving unique benefits in accomplishing efficient N₂/CH₄ separation. Traditional approaches, such as cryogenic distillation, pressure swing adsorption (PSA), and other conventional techniques, have been widely utilized for N₂/CH₄ separation. However, these methods suffer from several constraints, including high energy consumption, operational complexity, and limited selectivity, thereby underscoring the need for a more advanced separation technology.

Membrane technology, in particular, is an appealing alternative for addressing these issues due to its scalability, energy efficiency, and simplicity of integration with other processes. Further developments in material science and process optimization are required to produce robust and cost-effective N₂/CH₄ separation technologies that ensure the sustainable and efficient use of natural gas resources. Membrane separation has evolved as an energy-efficient and compact method of N₂/CH₄ separation [10]. This method employs polymeric or inorganic membranes that selectively permeate gases based on their size, solubility, and diffusivity [11]. Membrane technology is ideal for remote and offshore applications where modular systems are preferable. Nonetheless, conventional membranes frequently struggle with low selectivity for nitrogen and methane due to their similar molecular size [10]. Thus, further advancements in membrane materials are essential to enhance separation performance and economic feasibility. Although several membrane types have been investigated so far for potential membrane materials used in gas separation, however, studies reporting membranes with advanced properties capable of achieving high separation efficiency specifically for N₂/CH₄ gas mixtures appear to be lacking in this field.

As such, this review offers a comprehensive and critical synthesis of recent progress in membrane-based N₂/CH₄ separation, a process that is increasingly vital for natural gas purification and biogas upgrading. While membrane separation of gases has received broad attention, relatively few studies have tackled the specific challenges of N₂/CH₄ separation, particularly in terms of material design, performance trade-offs, and industrial viability. This article focuses on emerging membrane materials, particularly polymeric, inorganic and mixed matrix (MMMs) membrane incorporating metal-organic frameworks (MOFs), zeolites, and carbonized fillers and evaluates their selectivity, permeability, and structural compatibility. It further explores the convergence and divergence across studies, drawing attention to novel approaches and shared bottlenecks. In addition to performance evaluation, considerations of scalability, cost-effectiveness, durability, and sustainability are also critically examined to provide a forward-looking perspective on how current laboratory developments can be translated into practical separation technologies. Ultimately, the review aims to bridge material innovation with process engineering to guide future research and industrial implementation.

The separation of nitrogen (N₂) from methane (CH₄) is a critical process in several gas-related industries, including natural gas upgrading, biogas purification, and coal bed methane processing. In natural gas production, N₂ is often present as a diluent, reducing the energy content of the gas and increasing pipeline transport volumes. Raw natural gas streams can contain anywhere from 2 to 50% N₂, particularly in non-conventional reserves. Meanwhile, biogas and landfill gas streams can also contain notable amounts of N₂ due to air infiltration during collection. The industrial requirement for pipeline-quality natural gas typically mandates CH₄ purities above 95–98%, with N₂ content reduced to below 4%, depending on local standards and transmission company specifications. This necessitates a highly selective and energy-efficient separation method, as N₂ and CH₄ possess very similar kinetic diameters and boiling points, making their separation a challenging task. Table 1 further elaborates the physicochemical similarities between CH₄ and N₂ that present a major obstacle to membrane-based separation.

2.1 Cryogenic distillation

Cryogenic distillation is a well-established technology that is commonly used in gas separation procedures due to its effectiveness at separating gases with considerable boiling point differences. The technique uses the differential in volatility of molecules as a driving force for separation. The boiling points of N₂ and CH₄ are significantly different, with N₂ at −195.8 °C and CH₄ at −161.5 °C, which allows for effective separation [15]. This temperature difference enables the selective condensation and vaporization of components, making cryogenic distillation a dependable method for producing high-purity outputs. The distillation process can produce methane with a recovery rate of more than 98% while reducing nitrogen concentration in the methane stream to less than 1% [16]. While highly effective for large-scale industrial applications, cryogenic distillation is energy-intensive and costly, which calls for the adoption of cost-effective separation technology with high efficiency for large-scale separation processes.

2.2 Pressure swing adsorption (PSA)

Pressure swing adsorption (PSA) is a very efficient technique for gas separation based on the principles of adsorption and desorption, regulated by pressure variations. It separates specific gases by cycling between high and low pressure [17]. PSA uses adsorbents that capture nitrogen molecules better than methane molecules during CH₄/N₂ separation. The separation begins when the gas mixture flows into a packed column at pressurized conditions during the feed phase. The adsorbent binds N₂ molecules more strongly than CH₄ molecules because of their size differences, which results in higher CH₄ concentration in the output stream. After N₂ reaches maximum saturation in the adsorbent, the column pressure drops, allowing N₂ to partially leave the adsorbent reservoir. Despite the application of PSA in this regard, several challenges limit its application in achieving a highly efficient N₂/CH₄ separation. This includes the repeated pressurization and depressurization cycles that consume substantial energy, particularly at high throughput, reducing overall process efficiency. Consequently, the adoption of efficient separation technology with less energy requirement is of critical importance.

2.3 Membrane separation

Membrane separation works by allowing gas molecules to move through membranes at different speeds to separate one component from another [18]. The use of polymeric, inorganic, and MMMs for N₂/CH₄ separation was carefully analyzed and discussed in this present review to solve the existing separation problems. Polymeric membranes remain the preferred choice in gas separation due to their low production costs and easy expansion capabilities. The gas separation process of N₂/CH₄ also benefits from inorganic membranes due to their excellent thermal stability and chemical resistance, plus their extended lifespan. Inorganic membranes keep their shape better than polymeric membranes when exposed to extreme heat and pressure, which makes them ideal for natural gas processing [19]. Researchers build inorganic membranes from ceramic materials, zeolites, carbon molecular sieves, silica, and metal oxides. These membranes use three separation techniques: molecular size restriction, surface binding, and gas flow to separate CH₄ from N₂ despite their matching molecular dimensions.

Mixed matrix membranes unite polymeric and inorganic materials through the integration of inorganic fillers into a polymeric structure. This technique enhances polymeric membrane applications by adding inorganic components that provide high selectivity and stability while keeping the membranes flexible and affordable. The combination of MOFs, functionalized nanoparticles, or clay with EVA or Pebax matrices in MMMs produces superior N₂/CH₄ separation results. Recent studies demonstrate that hybrid membranes with functionalized nanofillers offer better separation performance and sustainability through increased permeability and selectivity.

Membrane gas separation is a complex process that depends on gas-material interactions, membrane morphology and physicochemical properties of the gases to be separated. In the especially difficult case of separating N₂ and CH_4, two non-polar gases with similar kinetic diameters, multiple transport mechanisms often operate in tandem. This section presents and discusses the most important mechanisms that drive separation performance in the framework of recent N₂/CH₄-selective membrane technologies, thus addressing the need for clearer mechanistic insights. Table 2 presents a summary of the key membrane separation mechanisms for N₂/CH₄ mixtures and representative material systems.

3.1 Solution-Diffusion mechanism (SDM)

The solution-diffusion model dominates in dense polymeric and MMMs, particularly where the polymer phase is continuous. Here, gas permeation follows a three-step process:

Step 1: Gas molecules dissolve into the membrane surface (solubility-dominated).

Step 2: Gas molecules diffuse through the polymer matrix (governed by free volume and segmental motion).

Step 3: Gas molecules desorb at the permeate side.

Mathematically, the permeability\(\:{P}_{i}\) of a species i is defined as:

where Di is the diffusion coefficient (kinetically influenced) and Si is the solubility coefficient (thermodynamically governed). Generally, CH₄ tend to exhibit higher solubility (due to greater polarizability as compared to N₂: 2.60 ų vs. 1.74 ų) but slightly lower diffusivity due to its larger kinetic diameter. Therefore, SDM-based membranes (e.g., SBS/Ni-MOF-74 [20], PBI/CuBTC [21]) often yield moderate CH₄/N₂ selectivity due to competing solubility-diffusivity effects.

3.2 Molecular sieving mechanism

This mechanism relies on the selective transport of molecules based on size exclusion through sub-nanometer pores. Ideal for rigid and microporous fillers (e.g., MOFs, zeolites), molecular sieving discriminates between gases with minute size differences. Molecules with smaller diameters permeate more preferentially. As observed from the reviewed studies, HOF-21/6FDA-DAM MMMs [22] achieve enhanced N₂ permeance via sieving through ~ 3.64 Å pores, selectively hindering CH₄ transport. Also, ZIF-8@VR/PAA MMMs [23] utilize carbonized surface defects to fine-tune diffusion channels, increasing CH₄ permeability by modifying access through the framework. Thus, molecular sieving is particularly relevant in MMMs engineered for N₂-selective membranes.

3.3 Facilitated transport mechanism

This mechanism involves the reversible interaction between target gas molecules and reactive carriers or binding sites within the membrane matrix, forming temporary complexes that migrate across the membrane [24]. Metal-rich MOFs (e.g., CoNi-DABCO in PDMS [25]) provide open metal sites, favouring CH₄ adsorption via van der Waals interactions. PVAm/CSZ membranes [Gu et al., 21] leverage Zn-rich carbonized MOF surfaces and amine-functional groups to selectively adsorb CH₄. This mechanism enhances selectivity by coupling transport with reversible adsorption, especially for CH₄-affinity-driven membranes.

3.4 Knudsen diffusion mechanism

In membranes with very fine pores (< 2 nm), where the mean free path of gas molecules exceeds the pore diameter, transport is dictated by molecular weight differences:

Where \(\:{D}_{K}\) is the Knudsen diffusion coefficient of speciesi(units: m²/s) and \(\:{M}_{i}\) is the Molar mass of gas species i (units: g/mol). CH₄ (MW = 16.04 g/mol) diffuses faster than N₂ (MW = 28.01 g/mol). This mechanism is not dominant in most MMMs discussed but may contribute marginally in ceramic supports or porous sub-layers.

4.1 Material structure and chemistry

Membrane materials’ structural and chemical properties directly affect their performance during N₂/CH₄ gas separation operations. These membrane characteristics directly affect free volume and chain rigidity and gas-polymer interactions, which determine gas permeability and selectivity. The contorted backbone structure of polymers of intrinsic microporosity (PIMs) produces high internal free volume, which leads to excellent permeability, but their undefined microvoids restrict their ability to separate gas molecules of similar size, such as N₂ and CH₄. The research by [23] showed that PIM-1 had high permeability, but its CH₄/N₂ selectivity remained low at 1.2–1.5 because of the disordered nature of its microporosity. To address this challenge, thermally rearranged (TR) polymers, especially polybenzoxazoles (TR-PBOs), derived from ortho-functionalized polyimides, have been studied for their enhanced rigidity and interchain interactions. These structural changes introduce well-defined micropores that facilitate size-selective transport. Beyond introducing micropores, these structural changes result in better-defined micropore channels that boost size-selective permeability.

According to [26] the incorporation of triptycene units into polymer membranes became possible by utilizing their three-dimensional shape-persistent structural properties. The ultrafine micropore network size achieved through this structural modification became the same size as gas molecules while improving molecular separation efficiency and transport speed. The combination of triptycene units in PIMs leads to optimal CH₄ permeabilities surpassing 1000 Barrer, together with selectivities that exceed 2.5 in N₂/CH₄ separation applications. The effective enhancement of membrane performance stems from backbone engineering and surface functionalization methods. The integration of polyacrylic acid and vapour-rubbed surface-carbonized ZIF-8 fillers (ZIF-8@VR) notably elevated CH₄/N₂ selectivity according to the study reported by Gu et al. [27]. The PAA/66.8%-ZIF-8@VR membrane with 800 nm thickness demonstrated a CH₄/N₂ selectivity of 3.12 and a CH₄ permeance of 1763 GPU, which surpassed the performance of both pure PAA and polysulfone membranes. The enhanced performance stemmed from the hierarchical pore structure of the filler combined with suitable CH₄ adsorption sites. The research demonstrates that membrane engineering by backbone rigidification and addition of triptycene frameworks or filling with functional porous materials leads to effective gas pathway modification. Membrane development for efficient N₂/CH₄ separation requires these innovations because they enable essential natural gas purification and nitrogen recovery processes.

4.2 Operating conditions

4.2.1 Effect of temperature

The performance of N₂/CH₄ separation across various membrane systems depends heavily on temperature due to its effect on both gas permeability and selectivity [28]. demonstrated that carbon molecular sieve (CMS) membranes experienced enhanced gas permeabilities when the temperature increased from 25 °C to 50 °C because of thermally activated diffusion. The membrane’s N₂/CH₄ selectivity decreased as temperature increased because sorption selectivity decreased primarily. The sorption coefficients and Langmuir hole-filling capacities decreased with temperature, which reduced the sorbed gas density within micropores, yet diffusion selectivity remained stable [29]. similarly observed that Ni-MOF-74 and SBS polymer-based MMMs showed increased CH₄ and N₂ permeabilities when operated between 25 °C and 55 °C. The increased mobility of polymer chains created additional free volume in the material. The temperature increase led to decreased CH₄/N₂ selectivity while CH₄ showed higher activation energy benefits compared to N₂ due to this selective change in permeability [30]. studied adsorption-selective carbon membranes to show that N₂ demonstrated high permeability at low temperatures because of strong adsorption interactions. The N₂/CH₄ selectivity experienced a significant decrease when the temperature reached 323 K, as N₂ adsorption weakened. The membrane’s enhanced activated diffusion mechanism enhanced both gas permeabilities and restored selectivity when operated at 373 K. Multiple experimental results demonstrate a consistent pattern where temperature increases enhance permeability through membrane flexibility but reduce separation effectiveness due to adsorption-based effects and diffusion behaviour, which become degraded.

4.2.2 Effect of pressure

The influence of feed pressure on N₂/CH₄ separation depends on membrane type and gas–membrane interaction characteristics, while pressure remains a vital factor for membrane optimization. The study reported by [31] explained pressure effects on PDMS and PEBA polymeric membranes through a pressure range from 0.30 to 0.80 MPa. CH₄ permeability increased with rising pressure because condensable gases like CH₄ entered the polymer structure and made it more flexible while creating additional free volume for gas diffusion. The less condensable nature of N₂ prevented it from interacting with the polymer, thus maintaining its permeability at a steady level. The different pressure responses between CH₄ and N₂ resulted in a small but significant improvement of N₂/CH₄ selectivity at elevated pressures.

The optimal membrane performance occurred at the highest pressure level combined with the lowest temperature setting, which demonstrates the importance of precise operational control. PEBA membranes maintained through PEBA 1074 showed the most promising performance for gas purification by outperforming PDMS-based membranes and achieving selective results for N₂/CH₄ separations. In addition [32], studied gas separation under low pressures to demonstrate how membrane-gas molecule interactions affect separation performance when strong binding occurs. The authors demonstrated that Cr-functionalized MXene membranes exhibited strong binding with N₂ gas molecules at a low pressure of 0.05 bar through Cr sites. The initial strong binding between Cr and N₂ molecules enhanced N₂ permeance, but excessive Cr loading resulted in decreased N₂ transport. The membrane’s strong binding interactions with N₂ molecules created a desorption problem, which prevented the gas molecules from moving forward.

The results from these various studies show that pressure increase improves CH₄ permeability with slight increases in selectivity but poor gas–membrane desorption kinetics become a limiting factor when working at low pressures. The successful separation of N₂ from CH₄ therefore depends on precise pressure control and membrane design, regardless of the material type, from flexible polymers to MXenes.

4.3 Presence of impurity

Membrane-based gas separation functions as a vital industrial technology that enables the separation of nitrogen (N₂) and methane (CH₄). Membrane performance suffers major degradation when impurities such as hydrogen sulfide (H₂S), sulfur dioxide (SO₂), and nitric oxide (NO) exist in the system because they cause permeability changes, selectivity degradation, and structural damage. Knowledge about these effects helps improve both membrane durability and efficiency when used in practical applications. Multiple effects occur when impurities interact with gas permeability. Research has shown that fresh membranes experienced a minor increase in nitrogen permeability when exposed to NO and H₂S conditions, possibly due to membrane structural plasticization [33]. The permeability reduction observed with SO₂ exposure indicates that competitive adsorption occurs because SO₂, with its higher critical temperature (431 K), displaces N₂ from available adsorption sites [12]. The permeability of MMMs decreased overall when exposed to impurities, yet H₂S caused the most substantial reduction despite competitive adsorption usually following critical temperature order (NO < H₂S < SO₂) [12]. Membrane performance depends heavily on impurity permeability as a key operational factor.

Studies revealed that SO₂ and NO permeability remained undetectable because their feed concentrations were too low. H₂S permeability exceeded N₂ and CH₄ permeability because of its elevated critical temperature and enhanced solubility properties [34]. It has been shown from the literature that membrane H₂S permeability increased when Metal-Organic Framework (MOF) or Porous Organic Polymer (POP) nanoparticles were added because they introduced new free volume spaces into the membrane structure. The Copper Benzene-1,3,5-Tricarboxylate (Cu-BTC) based MMM exhibited an outstanding performance by completely eliminating H₂S from the permeate stream, which indicated both irreversible adsorption and structural modifications in the MOF framework [35,36,37,38]. The permeability of membranes featuring different compositions experienced varying levels of deterioration after they encountered various impurity exposures. For example, pure Matrimid membranes maintained their stability in NO and H₂S environments for 60 days, but SO₂ exposure led to a slight decrease in membrane density through plasticization effects [34]. The permeability of POP-2-containing MMMs decreased by 50% during 3–6 weeks of testing, yet remained higher than pure Matrimid membranes, which indicates their potential use in flue gas applications. This indicates that exposure to impurities can cause substantial structural instability in membranes built with MOF materials. The permeability of Cu-BTC membranes decreased significantly because H₂S permanently bound to the membrane structure, which caused its collapse. The ZIF-8-based membranes demonstrated the worst structural deterioration because their permeability decreased to almost zero within seven days. The structural degradation of MOFs, combined with matrix blocking effects caused by impurities, resulted in a significant decrease in gas transport efficiency. The exposure to SO₂ and H₂S caused both membrane densification and crystallinity destruction, according to density analysis results [39, 40].

Membrane performance in gas separation processes experiences significant changes when impurities H₂S, SO₂, and NO are present. The mechanisms of competitive adsorption, together with plasticization and irreversible adsorption, modify both permeability and selectivity. The resistance of MMMs to contaminants varies, but continuous exposure to these substances eventually causes membrane deterioration and performance decline. Future research needs to create durable membrane materials that resist impurity damage to maintain stable performance in industrial gas separation systems.

4.4 Membrane morphology

Membrane morphology plays a crucial role in N₂/CH₄ gas separation as it determines both selectivity and permeability performance. Key morphological factors include pore size distribution, membrane thickness, and the presence of defects, all of which determine gas transport mechanisms. For instance [41], studied the synthesis of SAPO-34 membranes to separate N₂ from CH₄. The researchers achieved better N₂ permeance and selectivity through membrane synthesis in stainless steel autoclaves using controlled cooling methods which produced thinner membranes. While the membrane synthesized using water bath cooling techniques showed a combination of 2600 GPUs N₂ permeance and 7.4 selectivity levels demonstrating the impact of synthesis methods on membrane morphology and performance.

In another study [42], showed that membrane formation and quality depends heavily on the aspect ratio of SAPO-34 seed crystals which determines membrane morphology. The membrane thickness decreased from 6.6 mm to 2.4 mm when seed aspect ratio increased from 1 to 10 which resulted in membranes with smoother and more uniform packing and reduced defects. The controlled manipulation of seed geometry in combination with controlled seed deposition leads to morphological optimization of separation interfaces through which high-performance membranes are achieved. Furthermore, the activation temperature of ETS-4 membranes controls both membrane morphology and gas separation performance outcomes as reported by [43]. Membranes activated at 140 °C achieved the highest N₂ permeance and a permselectivity of 75.19 as a result of thermal treatment that optimizes membrane structure for better separation performance.

These studies collectively underscore the importance of precise control over membrane morphology to achieve optimal N₂/CH₄ separation performance. Tailoring factors such as pore structure, membrane thickness, and activation conditions can lead to significant improvements in selectivity and permeability, essential for efficient natural gas processing.​.

Nitrogen/Methane separation membranes include polymeric, inorganic, and MMMs, each with unique advantages for specialized applications. Polymeric membranes, including polyimides, poly(pyrrolones), and fluoropolymers such as Teflon^® AF and Hyflon^®, dominate the sector due to their simplicity of production, low cost, and balanced permeability and selectivity. These materials use variations in diffusivity and solubility between nitrogen and methane to separate [44, 45]. Inorganic membranes, such as ceramic and metallic materials, have high thermal and chemical durability, making them suitable for application in severe settings. These membranes separate gases via methods including molecular sieving and surface diffusion, making them ideal for high-temperature or chemically harsh environments. However, their high production costs and brittleness hinder commercial viability [46, 47]. MMMs combine the advantages of polymeric matrices with inorganic fillers such as zeolites, metal-organic frameworks (MOFs), and silica. These membranes improve separation performance by combining the flexibility and processability of polymers with the high selectivity and permeability of fillers. MMMs are becoming popular for N₂/CH₄ separations due to their adaptable features and superior gas separation performance [48, 49]. Choosing a membrane type for N₂/CH₄ separations involves balancing separation performance, operational needs, and economic feasibility.

Recent advancements have improved membrane performance, leading to more efficient N₂/CH₄ separation in industrial processes. Therefore, a comprehensive review on the use of membranes for N₂/CH₄ separation is necessary, particularly in natural gas processing and nitrogen rejection. Despite advancements in membrane technology, limited reviews have specifically addressed the progress, challenges, and future directions in this field. Hence, this review aims to consolidate existing knowledge, highlight recent developments, and identify research gaps. It is to be noted that some of the membranes considered in this review have a limited number of studies reported for their use in N₂/CH₄ gas separation applications; hence, the trends and insights obtained as a result of the collective review of these studies are solely based on the available data.

5.1 Polymeric membranes

The commercial utilization of polymeric membranes for gas mixture separation has been practiced since the late 1970 s, transforming gas separation operations by offering energy-efficient and environmentally sustainable alternatives to traditional methods such as adsorption and cryogenic distillation. The well-known trade-off between permeability and selectivity, embodied in Robeson’s upper-bound relationship, has limited the use of polymeric membranes for gas mixture separation. This connection, tested for a wide range of gas pairings including N₂/CH₄, reveals that achieving a high separation factor often comes at the expense of decreased permeability of the more permeable gas component [45]. Despite this constraint, current research has focused on addressing this trade-off using novel methodologies such as polymer blending, cross-linking, and filler integration to construct mixed-matrix membranes (MMM). Therefore, the following section presents a critical review of the latest advancements in polymeric membranes for N₂/CH₄ separation.

5.1.1 Polymers of intrinsic microporosity (PIMs)

According to a study reported by [50], different PIM-PI-OH membranes were synthesized using thermal imidization and one-pot polycondensation. The synthesis of PIM-PI-OH-2 and Copol-OH-(1–2) involved combining An-1 and An-2, while PIM-PI-OH-3 used DAR·2HCl as its diamine component. Structural and thermal analyses confirmed that the polymerization achieved its goal, thus validating the membrane characteristics. Table 3 shows the gas separation performance data for PIM membranes, which were prepared and treated differently for N₂/CH₄ separation. The researchers evaluated the membranes under three different conditions: their initial state, post-ethanol treatment, and after ageing. The membranes underwent structural swelling from ethanol treatment, which enhanced nitrogen permeability across all membranes while slightly decreasing N₂/CH₄ selectivity. The ageing process resulted in permeability reduction, but selectivity either stayed stable or improved slightly. The ethanol-treated O-PIM-PBO-1 membrane achieved the highest nitrogen permeability at 64 GPU yet demonstrated a low selectivity at 0.93, while the casted O-PIM-PI-OH-1 membrane showed the lowest permeability at 0.84 GPU with a relatively high selectivity at 1.22.

Additionally, in a study reported by Li et al. [51], PIM-1 was prepared by polycondensation of equimolar TFTPN and TTSBI in NMP with anhydrous K₂CO₃ under nitrogen, stirred, and heated at 60 °C for 24 h. The polymer obtained was precipitated in methanol, cleaned with dilute HCl and water, then dissolved in DCM and reprecipitated in methanol. It was finally dried under vacuum at 120^°C. Dense membranes containing 2 wt% of the PIM-1 in DCM were dissolved, filtered, and ring-cast at room temperature onto silicon wafers. The membranes were immersed in methanol after ~ 5 days of solvent evaporation and dried under vacuum at 120 °C. Films (~ 5 μm thick) were cross-linked by heating under vacuum at 250–300 °C, 0.5-2 days, to form flexible, defect-free membranes with insignificant weight loss. These thick membrane films were designated as PIM-temperature-duration (day), e.g. PIM-300-0.5d. PIM-1 membranes cross-linked at high temperature were more stable, decomposing more slowly, yielding > 65 wt% char at 750 °C, and containing > 99 wt% % gel after 0.5 d at 300 °C. As demonstrated in Fig. 1, thermal treatment of PIM-1 has a great impact on both permeability and selectivity. The original PIM-1b displays the highest permeability compared to the other samples. But, as thermal cross-linking continues, to the PIM-300 series and as the time is extended, both N₂ and CH₄ permeabilities decrease significantly, as does the free volume and chain mobility. It is noteworthy that although original membranes prefer the transport of CH₄ (N₂/CH₄ selectivity < 1), long thermal treatments switch this preference. The membranes sintered at 300 °C and 1.0 d or more are selective, >1, with the highest N₂/CH₄ selectivity of 1.32 in PIM-300-2.0d. This transition indicates that thermal cross-linking preferentially blocks CH₄ diffusion, presumably because of its larger kinetic diameter, increasing N₂ selectivity. In this way, the membrane is stabilized by controlled cross-linking, but also the separation performance is tuned.

Comparative Analysis of N₂ and CH₄ Permeability and Selectivity in Thermally Treated PIM-1 Membranes

Full size image

5.1.2 Comparative analysis and critical insight on N₂/CH₄ separation using functionalized and thermally treated PIM-Based membranes

Both studies investigate the modulation of membrane performance by physical/chemical treatment, ethanol swelling, and thermal cross-linking, pointing out the inherent trade-off between permeability and selectivity. The ethanol-treated PIM-PBO-based membranes achieved high nitrogen permeability (up to 64 GPU) but could not maintain selectivity, most samples having N₂/CH₄ selectivity of less than 1.35. This restricts their applicability in processes that need selective transport of nitrogen. Conversely, thermally treated PIM-1 membranes, which exhibit decreased permeability as a result of structural densification, showed increasing N₂/CH₄ selectivity as a function of thermal treatment. PIM-300-2.0d particularly achieved a selectivity of 1.32, which is better than the other regarding the separation efficiency.

The comparison has shown an important future research avenue, which entails the improvement of membrane microstructure to allow both high permeability and selectivity. While ethanol treatment is reversible and impacts selectivity negatively, thermal cross-linking offers a more durable route to tuning membrane performance. Additional measures in the form of hybrid modifications, the addition of nanofillers, or controlled cross-linking gradients may be needed to overcome the reduction in permeability that accompanies thermal densification. A compromise between the two studies could be achieved by hybrid methods that combine chemical functionalization (e.g., PBO groups) with controlled thermal cross-linking to achieve both higher free volume and maintained or improved selectivity. Considering industrial relevance, it is desirable that membranes preferential to CH₄ retention (i.e. selective to N₂) are desired in natural gas upgrading applications, and as such, the thermally treated PIM-1 series holds more promise than the PIM-PBO-based variants. Nevertheless, the high permeability of the former still holds merit, provided that the limitation of their selectivity can be circumvented. Further studies should investigate the long-term stability of thermally cross-linked PIMs under real operating conditions (e.g., pressure, contaminants), especially compared to swelling-prone ethanol-treated membranes.

5.1.3 Perfluorodioxolane

A study was carried out by [52], who investigated the novel use of amorphous perfluoropolymers, which are copolymers of perfluoro(2-methylene-4,5-dimethyl-1,3-dioxolane) and perfluoro(2-methylene-1,3-dioxolane) on gas permeances and selectivities. The study found that membranes made from some of these perfluoropolymers have superior gas separation (N₂/CH₄) performance compared to those prepared from the commercial perfluoropolymers (Teflon AF, Cytop, and Hyflon AD). The synthesized copolymers exhibited random polymerization (r₁ = 1.12, r₂ = 1.32), remaining amorphous, transparent, and fluorosoluble at 0–70 mol%. They showed low refractive indices (1.32–1.34), high Tgs (113–170 °C), and thermal stability (> 378 °C), outperforming conventional perfluorinated polymers.

The N₂/CH₄ permeation tests show that the newly developed perfluoropolymer membranes surpass commercial perfluoropolymers in terms of selectivity performance. The N₂/CH₄ selectivity of conventional polymer membranes stands at approximately 1, while commercial polydimethylsiloxane membranes show a preference for CH₄ over N₂ (< 1 N₂/CH₄ selectivity). The N₂/CH₄ selectivity of commercial Polysulfone and Matrimid membranes reaches only 1.1 and 1.3, respectively. The selectivity performance of Cytop membranes reaches approximately 3, while Copolymer D demonstrates a substantially higher selectivity, exceeding 5. The special characteristic of low methane solubility in perfluoropolymers enables the exploitation of their inherent N₂ diffusion selectivity. While Copolymer D’s selectivity is still lower than required for commercial applications, it represents progress toward developing viable N₂-selective membranes, which can reduce the cost of separation as mentioned earlier.

Further study on Perfluorodioxolane was carried out by [53] on the performance evaluation of Perfluorodioxolane copolymers on several gas mixtures, including N₂/CH₄. The study further reported that the N₂/CH₄ gas permeation data show that PFMMD–PFMD and PFMMD–CTFE copolymers surpass conventional perfluoropolymers in selectivity performance, although their behaviour patterns diverge because of their distinct molecular structures. The N₂/CH₄ selectivity of PFMMD–PFMD copolymers increases steadily from 4.3 to 6.0 as PFMD content rises from 25 to 60 mol% due to tighter chain packing and reduced free volume, which enhances diffusivity selectivity. The N₂/CH₄ selectivity of PFMMD–CTFE copolymers rises moderately from 4.5 to 5.5 when CTFE content increases from 37 to 57 mol%. PFMD-based copolymers produce higher N₂/CH₄ selectivity, although CTFE-based membranes deliver better gas permeability between them. PFMD incorporation makes the membranes suitable for natural gas purification applications as it increases their molecular sieving properties, while CTFE-based copolymers maintain ideal permeability and selectivity ratios for fast separation operations.

5.1.4 Comparative performance of perfluorodioxolane-based membranes

The studies by Okamoto et al. [51] and Fang et al. [52] reviewed show that perfluorodioxolane-based copolymers provide encouraging enhancements to commercial membranes in N₂/CH 4 separation. The two articles agree on the fact that these materials are more selective than conventional perfluoropolymers (e.g., Cytop, Hyflon AD) due to their low CH₄ solubility and enhanced N2 diffusivity. Remarkably, Okamoto and Fang reported N2/CH 4 selectivities of over 5 in Copolymer D and PFMMD PFMD systems, respectively, which is much higher than the ~ 1–3 of commercial membranes.

The difference, however, occurs in performance trade-offs: PFMD-based copolymers have greater selectivity (up to 6.0) at the expense of lower permeability, whereas CTFE-based copolymers have better permeability with moderate selectivity (4.5 to 5.5). These findings highlight one of the most important selectivity-permeability trade-offs and the impact of comonomer structure on transport properties.

These findings collectively suggest that the separation mechanism of the perfluorodioxolane-based membranes shifted from solubility-driven to a diffusivity-driven one. Contrary to traditional perfluoropolymers like Teflon AF, which usually prefer CH₄ because of its higher solubility, the membranes reviewed are highly selective to N₂. This tendency is mainly explained by the fact that CH₄ is not very soluble in the perfluorodioxolane matrix, and the role of solubility selectivity is reduced. Rather, it is diffusivity selectivity that becomes the dominant factor, as shown by the enhanced N₂/CH₄ selectivity with decreasing free volume due to more compact chain packing. The low CH₄ permeance and the relatively high N₂ permeance indicate that these membranes use size exclusion effects and differences in the kinetics of the gas molecules to increase the separation by molecular sieving. The behaviour is an encouraging step towards the creation of membranes capable of separating N₂ and CH₄ effectively, particularly in the natural gas upgrading process.

5.1.5 Polysulfone (PSF)

An investigation was carried out by [54] on the synergistic effect of adding a date pit powder on the physicochemical properties of a glassy Polysulfone polymer membrane for achieving the separation of nitrogen and methane. The researchers prepared PSF composite membranes through the dissolution of PSF in NMP and THF while adding de-oiled date pit powder (DP) at different weight concentrations (2–10 wt%) (DP-PSF-2 to DP-PSF-10) before casting the solution onto a glass plate. The membranes underwent controlled free-standing periods before being immersed in water, followed by methanol treatment and air-drying. The membrane characterization result confirmed that the date pit particles remain primarily at 20 μm with nonporous dense surfaces, defects at multiple sizes and more macrovoids when date pit content increased.

The study further stressed that the date pit-derived polysulfone (DP-PSF) membranes for N₂/CH₄ separation operate based on the solution–diffusion mechanism, where gas transport depends on solubility in the membrane and diffusivity through its matrix. Across all filler loadings (0–10 wt%), the membranes consistently showed N₂/CH₄ selectivity below 1, indicating a preference for CH₄ permeation. This suggests that CH₄, being larger but more condensable, diffuses more effectively due to favourable solubility interactions, while N₂ permeates less efficiently.

Further studies were carried out by Kluge et al. [55], who developed a thin separation layer polysulfone membrane with a sponge-like support structure as membrane morphology via wet-phase inversion. Also, the study was able to verify that the permeability and diffusion coefficient of a certain gas are not material-related properties but depend on the composition of the casting solution, amongst other factors. The SEM analysis showed that the PSU membranes consistently formed a two-layer structure: a dense gas separation layer and a sponge-like support. By reducing polymer concentration and adjusting solvent and non-solvent ratios, thinner separation layers were achieved. The most optimized membrane showed a defect-free separation layer with a thickness of several hundred nanometers, supported by a mechanically stable sponge-like structure. Large voids seen in earlier membranes were successfully minimized, improving structural integrity for high-pressure gas separation.

From the study, single-gas permeation experiments revealed that the symmetric PSU30_EIPS membrane had the best N₂/CH₄ selectivity of 1.09, and it was used as a reference because it had a dense and defect-free structure. PSU25_94:6 had the lowest selectivity (1.00), indicating almost equal permeation rates of N₂ and CH₄, whereas PSU30_80:20 had the highest selectivity of any asymmetric at 0.97. These values suggest little selectivity variation throughout the asymmetric membranes, which indicates small structural flaws or pinholes on the sub-micron scale that lower the efficiency of separation. N₂ and CH₄ permeance values were also less than 1 in all membranes, which is characteristic of dense structures and thin selective layers. Though permeabilities were somewhat higher in some asymmetric membranes, the calculated diffusion coefficients were lower than in the symmetric PSU30_EIPS, indicating that the polymer packing is denser in thin-film layers and thus limits gas diffusion and decreases selectivity. In general, none of the asymmetric membranes performed better than the symmetric PSU30_EIPS in N₂/CH₄ separation, which supports the importance of membrane integrity and layer thickness to selectivity. The findings indicate that the performance can be marginally enhanced by tuning the solvent systems and casting conditions, but the removal of defects is important in order to achieve greater selectivity.

Furthermore, Min et al. [56] fabricated a high-pressure hollow fibre polysulfone membrane module for gas permeation. The study reported the performance of the commercial polysulfone membrane, which was used as a barrier for several individual gases, including N₂ and CH₄. The Airrane Inc. membrane module performance on N₂/CH₄ gas separation shows that both gases have relatively low permeance of 1.20 GPU and 1.59 GPU of N₂ and CH₄, respectively. This slight variation shows that the membrane is not highly selective in the transport behaviour of N₂ and CH₄. The N₂/CH₄ selectivity of 0.75 is very low, which means that the membrane is not very selective to N₂/CH₄ and would not be able to discriminate between these two gases in a binary mixture, which is not ideal in processes where efficient N₂ removal in methane-rich streams is required.

The CH₄ behaviour was further assessed during the mixed gas separation tests with a simulated LNG-FPSO (Liquefied Natural Gas–Floating Production, Storage, and Offloading) gas composition. The concentration of CH₄ in the gas retentate rose with the stage cut, which indicated that methane was effectively retained on the feed side. This was, however, at the expense of reducing CH₄ recovery efficiency, meaning that more of the methane was not permeating through the membrane. Nitrogen did not form a significant part of the model gas and hence was not directly tested during these stages of performance. However, using the permeance data, it is possible to conclude that N₂ would have had a similar low-transport trend. To conclude, the Airrane module has poor N₂/CH₄ separation properties, and the permeance values are almost equal, and there is no strong selectivity advantage. Although the membrane can still be of some value in bulk gas separations, it is not suitable for applications where high-efficiency or precise N₂/CH₄ separation is needed without major system improvements or extra separation steps.

5.1.6 Critical comparative discussion of PSF membrane studies for N₂/CH₄ separation

Across the three studies, Adewole [53], Kluge et al. [54], and Min et al. [55], Polysulfone (PSF) membranes have N₂/CH₄ selectivities less than 1, favouring CH₄ transport (Table 4). This tendency is explained by the solubility-diffusion mechanism of PSF, in which the solubility and diffusivity of CH₄ prevail. Remarkably, every effort to alter PSF, through date-pit filler additions, structural tuning through solvent casting, or hollow-fibre forms, produces only minor modification to this methane-favouring baseline. Contrastingly, the addition of date pit powder by Adewole had a profound effect on permeability: low concentrations decreased permeability (because of tortuous pathways), whereas high concentrations increased it (through chain disruption and macrovoid formation). Nevertheless, such structural modifications did not change the basic selectivity that was methane-preferential at all filler concentrations. The asymmetric membranes of Kluge et al. demonstrated that even slight defects or alterations in filament thickness can further reduce selectivity, and the best symmetric layer had a maximum selectivity of only 1.09, demonstrating the limitations of morphological tuning. This trend was reproduced in practical operating conditions by the high-pressure PSF hollow fibres of Min et al., where CH₄ permeance was only slightly greater than N₂, further supporting the low selectivity of PSF to nitrogen-rich applications.

The experiments demonstrate a clear trade-off between selectivity and permeability. In the composites of Adewole, the permeability was enhanced with the higher loading of date-pits, but selectivity was not enhanced. Kluge et al. demonstrated that selectivity could be improved by thinning the selective layer, increasing permeance, but also creating pinholes that actually harmed selectivity. Min’s work underlines that high-pressure operation may further neutralize selectivity benefits, even as performance metrics remain within pipeline-quality ranges. All suggest a design constraint: without a fundamental change in membrane chemistry, performance gains remain modest and largely trade-off-limited.

One novel insight is that structural control or compositional changes cannot overcome the inherent bias of PSF to transport gases more readily than others; CH₄ will continue to permeate more readily than N₂ unless the polymer matrix itself is altered. The date-pit filler effects indicate that although fillers create macrovoids and modify the permeability, they have minimal control over gas selectivity. PSF’s main utility may thus lie not in achieving high N₂ selectivity, but potentially in bulk CH₄ separation or pre-treatment roles, where throughput is more critical than N₂ rejection.

Future studies should not just focus on morphological optimization to realize the potential of PSF in N₂/CH₄ separation, as this may only contribute little to no positive effect on the membrane performance. Chemical functionalization or mixed-matrix inclusion (e.g., incorporation of N₂-selective fillers such as ETS-4, zeolites, or MOFs) can introduce the molecular sieving necessary to achieve N₂ selectivity. The coupling of the mechanical strength of PSF with high-performance separation may be possible with hybrid composite membranes (PSF + selective inorganic phases). And finally, to know the performance in the real world, long-term testing in mixed-gas streams at different pressure and temperature is necessary.

5.2 Inorganic membranes

Inorganic membranes are strong and thermally stable materials which are largely used in gas separation because of their high chemical resistance and adjustable pore structure. In N₂/CH₄ separation, inorganic membranes provide an alternative to polymeric membranes, especially in high-pressure or high-temperature applications where conventional membranes can fail. Their selective transport properties may be designed to preferentially transport either nitrogen or methane, allowing efficient separation critical to natural gas upgrading and nitrogen rejection processes.

5.2.1 SAPO-34

A study was carried out by [42] group who synthesized highly crystalline SAPO-34 seeds with varying average aspect ratio (AR) and different levels of silicon doping to achieve a separation of N₂/CH₄ gas mixture. The synthesis of SAPO-34 membranes (M1 to M4) occurred through secondary growth on porous tubular α-Al₂O₃ supports using a required gel composition under controlled conditions. The inner surface of tubular supports received SAPO-34 seeds (S1 to S4) through a rubbing process to initiate membrane growth. Membranes M1, M2, M3, and M4 were synthesized through the use of seeds S1, S2, S3, and S4, which produced membranes with average thicknesses of 6.6 mm, 4.5 mm, 2.4 mm, and 1.8 mm. The XRD patterns confirmed that the synthesized SAPO-34 seeds displayed high crystallinity and phase purity. The SEM imaging revealed uniform dimensions for seeds with lower ARs but higher AR seeds showed broader size distributions and achieved extreme thinness reaching 0.015 mm. All seeds preserved their high crystallinity levels even though their thickness decreased.

The SAPO-34 membrane performance analysis demonstrated that membrane characteristics were deeply affected by the seed aspect ratio, through which membrane thickness was controlled. Also, the permeance and selectivity were enhanced. The membranes became thinner when the seed aspect ratio increased, which resulted in better separation performance. Membranes manufactured with seeds having elevated AR values exhibited superior packing efficiency, which produced smooth, thin layers with few imperfections. Membrane M3 produced with S3 seeds (AR ~ 10) demonstrated the best N₂/CH₄ selectivity of 11.1 ± 0.46 and N₂ permeance of 857 ± 44.8 GPU while surpassing all other configurations. The selectivity of M4 (AR ~ 20) decreased due to the seeds that entered support pores, which demonstrates the need for precise seed and support selection. The study shows that membranes with higher average aspect ratio seeds in SAPO-34 structures achieve better N₂ permeance alongside high N₂/CH₄ selectivity when silicon doping (M5) is applied to boost permeance without affecting selectivity. The data shows that silicon-doped SAPO-34 has promising applications for industrial CH₄/N₂ separation. However, there is insufficient understanding of gas transport behavior due to an absence of detailed structural analysis along with long term stability tests and description of mechanisms. Future research needs to include detailed structural analyses, stability tests and theoretical models to optimize synthesis processes and improve SAPO-34 membrane implementation in natural gas and biogas upgrading.

Building on the previous work by [42], another study was carried out by [57] who further optimized the synthesis conditions for N₂/CH₄ separation. The study reported a similar procedure for preparing SAPO-34 membrane with that reported by [42] at 50/50 pre-mixed N₂/CH₄ gas. But the order of chemical additions in this study was optimized to enhance monomer dispersion in the colloidal solution. Crystallization was carried out at 230 °C using water contents of 150, 200, 250, 300, and 400 to prevent gel phase separation and improve membrane quality. XRD results indicated that SAPO-34 exhibited the characteristic Chabazite topology throughout the examined water content range from 150 to 400 without showing any indication of phase separation. SEM imaging showed that the membrane surfaces contained micron-range zeolite crystals which grew well together while cross-sectional views demonstrated that the zeolite layer became thinner from 6.8 μm to 2.7 μm when water content increased.

It was observed from the study that the N₂/CH₄ gas permeation performance of SAPO-34 membranes strongly depends on the amount of water used during membrane fabrication. Table 5 demonstrates that higher water content during preparation produced thinner membranes that resulted in better N₂ permeance. Membranes prepared with 250 water content reached their highest N₂ permeance of 1200 GPU while maintaining a selectivity of 6.5. The highest selectivity value of 7.2 was achieved when the water content reached 300. The membranes developed defects at a water content of 400 which prevented them from sustaining pressure during permeation tests. This was attributed to the presence of large intergrown zeolite crystals, leading to defective intercrystalline boundaries. Hence, the optimal performance was achieved at 250 and 300 water contents, balancing high permeance and selectivity.

The research shows how gel dilution techniques improve N₂ permeance and selectivity in SAPO-34 membranes to reach their highest reported N₂ permeance level among the majority of inorganic membranes. Notably, the membranes exhibited remarkable reproducibility, as indicated by consistent N₂/CH₄ selectivities (~ 7.1–7.2) and N₂ permeance values (830–900 GPU) However, the researchers did not evaluate the stability or long-term performance of these membranes when operating under industrial conditions. The research did not perform a thorough investigation of defect formation processes at elevated water content levels. A study should explore both the post-synthetic modification methods and optimized crystallization conditions to limit the number of defects while also conducting prolonged stability tests for industrial membrane assessments.

In an effort to improve the process condition and selectivity of nitrogen gas in its mixture with methane [41], carried a study on the impact of autoclave materials and cooling methods. The study demonstrates that stainless steel autoclaves with controlled cooling significantly enhance selectivity (up to 8.6) and permeance (up to 2600 GPUs). These findings underscore the importance of synthesis conditions in tuning the SAPO-34 membrane performance. The gas permeation results demonstrate that stainless steel autoclaves produce better membrane performance than Teflon autoclaves when synthesizing SAPO-34 membranes through various cooling methods. The stainless steel autoclaves produced membranes with higher N₂ permeance by 2.2, 1.2, and 1.5 times compared to Teflon autoclaves when using water bath, ice/water bath, and flowing water cooling methods, respectively, due to stainless steel’s faster heat flux rates, which resulted in thinner membranes with fewer defects. The combination of ice/water bath cooling produced the highest N₂/CH₄ selectivity value of 8.6 for stainless steel membranes, while flowing water cooling achieved 9.2 for Teflon membranes. The separation indices (π), which reflect membrane performance and reproducibility, were generally twice as high for stainless steel membranes compared to Teflon ones, emphasizing the advantage of stainless steel in promoting better gas separation performance. The summarized performance results and insights are presented in Table 6.

5.2.2 Critical comparative analysis of SAPO-34 membranes for N₂/CH₄ separation

Across the reviewed studies, a shared focus on optimizing SAPO-34-based inorganic membranes for N₂/CH₄ separation is evident. All three studies used secondary growth methods on porous 2-Al₂O₃ supports, SAPO-34 as the zeolitic framework, and adjusted synthesis parameters (e.g. seed aspect ratio, gel composition, crystallization conditions) to control membrane microstructure. One of the main points of convergence of the studies is that there is a direct relationship between the decreased membrane thickness and enhanced gas permeance, in this case, N₂. As an example, membrane M3 and M5 of Huang et al. with 2.4 μm-thickness had N₂ permeances of 857 and 1200 GPU (1 GPU = 3.348 × 10⁻¹⁰ molm-² s-¹Pa⁻¹), respectively, and selectivities greater than 11. Likewise, the membranes prepared by Zong et al. using 250 300 water content attained the highest N₂ permeance (1200 GPU) and selectivity (7.2). This was further advanced by the work of Zong & Carreon, which showed ultrathin membranes (~ 1.8 μm) to have remarkable permeance up to 2600 GPU, particularly when stainless steel autoclaves and ice-water cooling were used.

While the goals align, the strategic focus varies. Huang et al. optimized seed morphology and silicon doping to control crystallinity and membrane density, Zong et al. focused on gel dilution and water content optimization, with emphasis on the importance of precursor solution dynamics, and Zong & Carreon studied autoclave material and post-synthesis cooling, a relatively unexplored parameter. These differences are reflected in the outcomes. As an illustration, the silicon-doped membrane (M5) developed by Huang et al. had high selectivity and increased permeance, indicating that the chemical composition adjustment can overcome the trade-off. Conversely, Zong & Carreon showed that membrane morphology and functional properties can be drastically improved by thermal gradient control during crystallization. Interestingly, their reported N2 permeance (2600 GPU) is higher than the other studies, indicating that process engineering may be better than material changes.

A novel interpretation of this comprehensive review is that while most MMMs or MOF-based polymeric membranes struggle to balance high selectivity and permeability, due to their inherent trade-offs, SAPO-34-based inorganic membranes can avoid this to a certain degree by using carefully controlled microstructures, modulation of seed and gel chemistry and optimization of thermal processes. These methods give selectivities greater than 5 and N₂ permeances greater than 1000 GPU, which, as indicated on the Robeson plot, place many data points above or at the edge of the Robeson upper bound of N₂/CH₄. This is in stark contrast to the typical polymeric membranes, which occupy the lower-left quadrant of the plot, where log₁₀ (P) < 2.5 and log₁₀ (α) < 0.7. Furthermore, this synthesis proposes an encouraging trend of N₂-selective membrane development, a direction less emphasized historically. The majority of conventional membranes are CH₄-selective, due to their greater commercial value, yet N₂-selective membranes can streamline natural gas upgrading by lowering the cost of recompression.

The data points lotted with the Robeson upper bound (Fig. 2) indicate that the membranes prepared by Huang et al. and Zong & Carreon consistently challenged or exceeded those of the upper bound, Zong et al. have achieved parity with the upper bound at optimal water content, and the Polymer-based membranes are below the upper bound, indicating their fundamental limitations in high N₂ selectivity with permeability. This confirms that more sophisticated inorganic approaches, particularly those involving SAPO-34, provide a pathway to surpass the conventional permeability-selectivity trade-off curve.

Four primary directions should be the focus of future research. Firstly, the use of SAPO-34-based MMMs through the incorporation of these inorganic fillers into flexible and robust polymeric matrices to achieve a balance between processability and top-tier separation performance. Secondly, the development of effective gas transport models specific to N₂-selective pathways in SAPO-34, based on molecular dynamics and adsorption diffusion simulations to optimize membrane structure, should be seriously considered. Thirdly, realistic industrial conditions such as impurity resistance, thermal cycling, and mechanical stress testing should be performed over long periods. Lastly, future research should transition from lab-scale fabrication to pilot-scale module development, evaluating the system-level energy savings in the case of natural gas upgrading or biogas purification.

Comparative Performance of SAPO-34 Membranes for N₂/CH₄ Separation

Full size image

5.2.3 SSZ-13

SSZ-13 zeolite membranes are known for their excellent molecular sieving properties, making them highly effective for gas separation applications. As such, several studies have been carried out on the application of this class of membrane material for achieving high selectivity of nitrogen gas in natural gas purification. One of such study was carried out by [58]. The study reported the performance of SSZ-13 membranes on enhanced separation processes for N₂/CH₄ separations. The SSZ-13 membrane synthesis was carried out using asymmetric α-alumina supports, and the outer surface was coated with seed crystals. The membrane gel was prepared with a modified composition and hydrothermally treated, which was then calcined in pure oxygen. The XRD analysis demonstrated that all SSZ-13 samples contained pure chabazite phase with a high level of crystallinity, while the SEM images showed continuous zeolite films across all membranes. The membrane reached a thickness of 6.0 μm while an intermediate α-alumina layer stopped zeolite from penetrating deeply into the support.

The SSZ-13 membranes have a good potential for N₂/CH₄ separation as shown by the performance data in Table 7, mainly because of the molecular sieving effect, which prefers nitrogen to methane. In single gas tests, the N₂ permeance was 203 GPU and the selectivity was high (13), which implies that the two gases were well discriminated because of the smaller kinetic diameter of nitrogen. The membrane performed solidly in dry equimolar mixtures with an N₂ permeance of 209 GPU and a selectivity of 6.2, and was stable and consistent with single-gas measurements. At 323 K under humid conditions (99.9% RH), the permeances of both N₂ and CH₄ dropped by a large amount (to ~ 33% and ~ 54% of the initial values, respectively), resulting in a lower selectivity of 3.85. Performance was, however, restored after drying, indicating that the membrane is reversibly influenced by water vapour. The selectivity further rose to 5.14 at a higher temperature of 373 K, indicating that desorption of water at high temperatures enhances the efficiency of separation. Also, a membrane of 6.2 m thickness reported the highest N₂ permeance of 253GPU and a selectivity of 9.5, which confirms the applicability of the SSZ-13 structure in gas separation processes. In general, these membranes have high nitrogen permeability, moderate to high selectivity, and good stability, even in harsh humid environments, and thus are promising toward industrial-scale natural gas upgrading in which N₂ removal is needed.

Building on previous findings of SSZ-13 membrane stability and selectivity under humid conditions [59], carried out a study aimed to enhance N₂/CH₄ separation efficiency by using modified membrane gel compositions and nano-sized ball-milled SSZ-13 for secondary growth. The SSZ-13 membranes used in this study were prepared following the same procedure as highlighted by [58] but with a different composition of SiO₂/Al₂O₃ ratio in the membrane gel from 40 to 200. The characterization result of the membrane confirmed that the Chabazite phase structure remained intact after ball milling, with high-intensity peaks indicating high purity and reduced SSZ-13 crystal size from approximately 3 μm to 300 nm. The research investigates SSZ-13 membrane performance for N₂/CH₄ separation by studying temperature and pressure effects on permeance and selectivity. The membrane’s N₂/CH₄ selectivity drops from 9 to 6 when temperature increases, due to nitrogen permeance reduction while methane permeance stays constant. The lower kinetic diameter of nitrogen at 0.367 nm compared to methane with 0.380 nm results in faster diffusion, which proves that molecular sieving drives the separation process instead of preferential adsorption. The membrane’s nitrogen permeance decreases when pressure rises, while selectivity decreases from 13 to 9. The membrane’s minimal defects result in low methane permeance, making it N₂-selective. These observations align with trends noted in the previous study, highlighting the importance of molecular sieving in N₂/CH₄ separation.

Another study was carried out by [60] who synthesized SSZ-13 by a novel seeded-gel approach for the first time. The study optimized SSZ-13 membrane synthesis parameters to improve their ability to separate N₂ from CH₄ molecules. Unlike previous studies that primarily focused on general SSZ-13 membrane fabrication, this work explores the influence of changing seed levels, hydrothermal synthesis temperatures and various calcination atmosphere (air, oxygen, and ozone/oxygen) on membrane structure, defect minimization, and selectivity improvement. XRD analysis proved the complete removal of OSDA from the material along with a higher degree of crystallinity through calcination. SEM imaging showed crystals shrank from 4 μm to 200 nm and produced a well-intergrown membrane that was 1.2 μm thick. Through seeded-gel synthesis a more uniform thin membrane was formed than through rub-coating and it showed improved separation functions.

The optimal seed concentration (0.7 mg seeds/g gel) resulted in the highest N₂/CH₄ selectivity (13.5), whereas deviations led to lower selectivities due to defect formation. The study further reported that the calcination in an ozone atmosphere at 473 K for 48 h provided superior selectivity (11.6) compared to oxygen or air at 673 K (6.2 and 2.9, respectively), as higher temperatures induced defects. Prolonged ozone calcination (beyond 48 h) had minimal additional impact on permeation. The membranes exhibited a high N₂ permeance of 2500 GPU, demonstrating their efficiency for gas separation. Increasing pressure from 0.2 MPa to 2.6 MPa decreases N₂/CH₄ selectivity from 5.2 to 1.7, indicating defect-related non-selective transport. Similarly, raising the temperature from 25 °C to 200 °C reduces N₂ permeance from 161 to 47.8 GPU and selectivity from 5.2 to 1.8, likely due to enhanced diffusion and reduced CH₄ adsorption. Although specific values for flow rate are not provided, higher flow rates typically lower residence time, reducing selectivity. It was observed that the present study provides first-time experimental evidence of zeolite membrane performance for N₂/CH₄ separation at high pressures (> 2.5 MPa), improving understanding of temperature, pressure, and flow rate effects on gas separation.

As observed from the previous studies, the development and fundamental research on N₂-selective zeolite membranes is still challenging to scale up; hence, a study was carried out by [61] on scaling-up the production of highly N₂-selective SSZ-13 membranes for the first time. SSZ-13 membranes were fabricated on a 50 cm long alumina support via secondary growth with vacuum seeding, ensuring denser seed layers. After synthesis, membranes were dried, calcined, and stored for use. The SSZ-13 membranes underwent SEM and XRD analysis, which revealed a continuous defect-free compact structure with a membrane thickness between 3.5 and 3.9 μm. The vacuum seeding technique produced an even seed distribution to solve the problems of dip-coating that resulted in membrane defects and poor crystal integration. The use of an oil-bath heating system resulted in uniform membrane quality and was further improved through optimized gel ageing.

The study emphasized that these SSZ-13 membrane showed successful N₂/CH₄ separation with N₂ molecules permeating faster because of molecular sieving effects. The N₂ permeance reached 406 GPU at 298 K while showing a decreasing trend with both temperature and pressure drop. CH4 permeance showed no change in response to temperature variations but increased as pressure drop levels rose because of defect pores, which is consistent with previous similar studies. The membrane selectivity decreases from 13.3 to 6.2 when temperature increases and from 13.3 to 9.2 when the pressure drop increases. The separation performance decreases when feed flow rates decrease because of concentration polarization effects. The permeance and selectivity of N₂ decreased when the membrane was exposed to impurity gases, ethane and propane, with propane showing a more pronounced effect. The membrane functioned reliably even while exposed to impurity concentrations reaching 10% ethane and 5% propane.

The research shows that SSZ-13 membranes with an area of 180 cm² can be consistently produced through secondary growth while maintaining high N₂/CH₄ selectivity at 13.3 ± 0.5 and notable N₂ permeance. However, both selectivity and permeance decrease with increasing temperature and pressure drop, likely due to concentration polarization (CP) effects. The research indicates that SSZ-13 membranes present outstanding potential for natural gas purification as they demonstrate stability alongside scalability and resistance to impurities. Further optimization could enhance performance under varying operational conditions.

5.2.4 Comparative summary of SSZ-13 membrane studies for N₂/CH₄ separation

The studies reviewed confirm the outstanding efficiency of SSZ-13 zeolite membranes in the separation application of N₂/CH₄, which is mainly attributed to molecular sieving. This process is explained by the fact that the kinetic diameter of N₂ (0.364 nm) is smaller than that of CH₄ (0.380 nm) and thus N₂ can diffuse selectively through the zeolitic pores. In the studies reviewed, the effect of raising temperature or pressure is reported to affect the N₂ permeance and the total N₂/CH₄ selectivity. In particular, high temperatures decrease the adsorption of nitrogen on the membrane, and high pressure increases non-selective transport routes or leads to saturation. It is interesting to note that methane permeance is not strongly dependent on the operational conditions, which proves that the diffusion through SSZ-13 is not adsorption-kinetic-controlled, but rather size-selective.

A convergence in findings is observed across all studies regarding the sensitivity of N₂/CH₄ separation performance to temperature, pressure and feed composition (Fig. 3). For instance, Song et al. and Li et al. both indicate a declining selectivity with rising temperature and pressure, while Zhang et al. also confirm this tendency in the case of impurity-containing feeds. Moreover, the membranes exhibit moderate nitrogen permeance based on synthesis route and membrane thickness, and exhibit good performance stability during short-term testing (< 60 h). There are, however, a number of divergences between the studies. This includes membrane manufacturing methods and scalability potential, among others. Although most research papers have carried out small-scale membrane tests, Zhang et al. were able to scale up the membrane synthesis to a 180 cm² surface area through vacuum seeding and oil-bath heating, and attained a high N₂/CH₄ selectivity of 13.3. This development helps to fill the laboratory and industrial gap.

This review provides several novel insights not explicitly emphasized in the original studies. First, there is a common trend between the N₂/CH₄ selectivity decrease and the temperature and pressure increase, irrespective of the synthesis method. Second, defect control, especially by means of ozone calcination and seeded-gel synthesis, is more determinant of membrane selectivity than composition changes or support type. Third, the use of impurity gases (ethane and propane) by Zhang et al. emphasizes the role of impurity tolerance in practical applications, which was not much addressed in previous research. Finally, although short-term stability tests show promising outcomes, the literature lacks a considerable amount of information on long-term stability testing (more than 60 h), membrane fouling, and operational durability in industrial settings. Collectively, this analysis indicates that the optimal SSZ-13 membrane for industrial-scale nitrogen separation will incorporate the fine defect control of the seeded-gel method of Li et al. with the scalable fabrication method of Zhang et al. The hybrid approach provides a route to the fabrication of thin, defect-free, and high-performance membranes with reproducible and scalable results.

Furthermore, the Robeson upper bound points out a trade-off in polymeric membranes. The higher the permeability, the lower the selectivity. In N₂/CH₄ separation, the most promising polymer membranes have selectivities of 4.4 at low permeabilities (e.g., 0.31 Barrer), and fall below 2 at permeabilities above 1000 Barrer. Conversely, SSZ-13 zeolite membranes that have been reported in recent literature always exceed this limit. Even under humid and high-pressure conditions, selectivities of 13.5 and N₂ permeances of 2500 GPU were realized. Although direct comparison to the Robeson plot is limited due to the lack of membrane thickness data in the original studies, which precludes permeability (Barrer) conversion, the elevated selectivity values and performance trends suggest that SSZ-13 membranes potentially exceed the upper bound of conventional polymeric membranes. This affirms the potential of SSZ-13 as a high-performance substitute for energy-efficient N₂/CH₄ separation, which addresses the drawbacks of polymer-based membranes.

Future research on SSZ-13 membranes should focus on overcoming the key challenges to allow industrial use, especially by optimizing the synthesis procedures to remove nano-scale defects that reduce selectivity in practice. The significant reduction in N₂/CH₄ selectivity at higher temperatures and pressures requires structural improvements like composite or hybrid structures. Durability remains a key gap, with no study exceeding 60 h of testing, necessitating long-term and cyclic stability evaluations. Also, the tolerance to impurities should be strictly tested with practical gas mixtures that include CO₂, water vapour, and higher hydrocarbons, other than ethane and propane. Lastly, techno-economic studies are necessary to confirm lifecycle cost-effectiveness and aid commercialization, particularly for scaled-up fabrication processes such as vacuum seeding and oil-bath crystallization.

Venn diagram illustrating the convergence and divergence among four key studies on SSZ-13 membranes based on synthesis methods, operational conditions, and functional improvements

Full size image

5.2.5 AlPO-18

In a study reported by [62], a demonstration was done on the separation ability of AlPO-18 membranes for N₂/CH₄ gas mixtures. These AlPO-18 membranes displayed unprecedented N₂ permeances at moderate N₂/CH₄ separation selectivities. The AlPO-18 seed crystals were prepared and a synthesis gel with a molar ratio of 1.0 Al₂O₃:1 P₂O₅:1.8 TEAOH:120 H₂O was used for the growth of the membrane. AEI topology was confirmed by XRD and SEM revealed a dense layer of 2.4 μm with well-intergrown crystals. Moderate selectivity (4.6) and high N₂ permeance (3076 GPU) were found. Separation index (pi) which was measured in the range of 0.11–0.25 mol/(m²s) showed consistent results in performance. CH₄ was adsorbed 2.7 times more than N₂ since competitive adsorption favored CH₄. In breakthrough experiments N₂ was found to pass through the membrane faster than CH₄, which confirmed that the primary separation mechanism was the difference in diffusivities. The AlPO-18 membranes are more effective in N₂ separation due to higher diffusivity, but the separation efficiency may decline at high pressures or in presence of impurities.

Further studies were carried out by [63] on the application of AIPO-18 as a potential membrane for N₂/CH₄ separation. The study focused on the preparation of the AlPO-18 crystals with the sole DIPEA template and investigated the use of AlPO-18 membrane for single gas permeation of N₂/CH₄ gas mixture. In addition, the influences of temperature and pressure on the separation were studied. The AlPO-18 membranes were synthesized using seed crystals (1Al₂O₃:1P₂O₅:2DIPEA:60 H₂O).

A synthetic mixture (1Al₂O₃:1P₂O₅:mDIPEA:120 H₂O:4IPA) was stirred for 24 h, and uncalcined seeds were coated onto mullite supports before crystallization. The XRD analysis showed that increasing the amount of DIPEA content caused a reduction in the impurities, thereby enhancing the membrane crystallinity and the separation efficiency. SEM images also demonstrated the importance of microporous crystal integrity in attaining reproducibility and performance. The authors of the research assessed the permeation characteristics of AlPO-18 membranes (M-1, M-2, and M-3) synthesized at various DIPEA/Al₂O₃ molar ratios. The membranes exhibited different N₂/CH₄ selectivity since they had different structural properties. The selectivities of N₂/CH₄ of M-1 and M-3 were 4.9 and 6.8, respectively, since the N₂ molecules passed through the membrane pores more quickly than the CH₄ molecules due to their smaller kinetic diameter. The M-2 membrane structure had defects that allowed CH₄ to permeate through with ease, and therefore its selectivity value was low at 0.9. The findings validate that molecular sieving is a major factor in N₂/CH₄ separation, yet defects have a major influence on the mechanism of gas transport. An increase in temperature (303 K to 423 K) resulted in a decrease in N₂ permeance, but CH₄ permeance did not change, lowering selectivity. The change of pressure between 0.2 MPa and 1.0 MPa did not make much difference, but CH₄ permeance was a little higher because of defect-controlled viscous flow.

Furthermore, Wang et al. [64], reported that the AlPO-18 membrane (M2) synthesized under optimized conditions, exhibited improved microstructure with a continuous 8 μm thick zeolite layer and minimal crystal intrusion into the support pores. This morphological improvement led to improved gas permeation. In particular, at 298 K and 0.2 MPa, the permeance of N₂ and CH₄ were 41.78 GPU and 8.82 GPU, respectively, which resulted in an ideal N₂/CH₄ selectivity of about 4.7. This moderate selectivity shows that the AlPO-18 membrane is able to differentiate between N₂ and CH₄, although they have similar kinetic diameters (0.364 nm and 0.38 nm, respectively). The increased N₂ permeance is probably due to its slightly smaller size and greater diffusivity, and not due to the selectivity of adsorption of the two gases on the AlPO-18 framework, since both gases are weakly adsorbed on the framework. The much reduced CH₄ permeance implies that it has a higher diffusion resistance to the AlPO-18 pore dimension (~ 0.38 nm) and therefore increases the size-exclusion effects. These findings suggest that the AlPO-18 membrane transport in this regard is dominated by molecular sieving. The selectivity is however lower than that of strongly adsorbing gas pairs, which supports the view that adsorption strength and size exclusion are key parameters in high performance separations in zeolitic membranes.

5.2.6 Critical comparative analysis and emerging trends in AlPO-18 Membrane-Based N₂/CH₄ separation

The comparative analysis of AlPO-18 membrane studies reveals several novel insights that extend beyond the scope of the original publications. One major trend is the strong influence of synthesis conditions, particularly the choice and concentration of organic templates such as TEAOH and DIPEA, on membrane performance. These parameters significantly affect defect density and microstructural integrity, which in turn control the trade-off between selectivity and permeance. This suggests that defect engineering, rather than topology alone, plays a central role in determining the membrane’s separation behaviour. Additionally, while it is commonly assumed, as observed from previous studies, that thinner membranes are always advantageous for enhancing permeance, the reviewed studies reveal a more nuanced reality. For instance, thicker membranes (e.g., 8 μm) with higher structural uniformity may have a higher selectivity due to lower defect intrusion, which indicates a non-linear correlation between thickness and performance.

Another key insight is that there is a lack of consistency in the performance of membranes with identical AEI frameworks. Although membranes had identical crystalline topology, N₂/CH₄ selectivity varied widely, supporting that structural order, intercrystalline connectivity and defect-free growth are equally important as chemical composition. Moreover, the review indicates that the N₂/CH₄ separation in AlPO-18 membranes is largely governed by the difference in diffusivity, as opposed to adsorption, since neither of the gases interacts strongly with the AlPO-18 framework. This implied that future modification should be focused on precision in pore size and geometry rather than on enhancing adsorption affinity. Notably, the presence of even minor defects, such as those observed in one of Liu et al.’s membranes, can entirely reverse selectivity, allowing CH₄ to permeate faster than N₂. This highlights the importance of defect detection and minimization in synthesis and fabrication.

Finally, the studies considered single and mixed-gas performances, but none of them incorporated dual-mode or multi-mechanism transport models to comprehensively predict and describe permeation trends. This review has highlighted this gap and proposed that the inclusion of transport modelling, which takes into consideration both diffusion and adsorption, may lead to a more complete picture of gas separation behaviour, especially under industrial conditions. Altogether, these insights not only help to explain the main structure-performance relationships in AlPO-18 membranes but also give a strategic direction to optimize the fabrication process, increase reliability, and customize membrane materials to practical N₂/CH₄ separation processes. Table 8 provides a summary of these major insights.

5.2.7 ETS-4

ETS-4 is a microporous titanosilicate molecular sieve with thermally tunable pore sizes (~ 3.6 Å), enabling molecular sieving between nitrogen and methane. Its rigid framework selectively allows N₂ (3.64 Å) to permeate while restricting CH₄ (3.80 Å), making it effective for nitrogen rejection in natural gas purification. Unlike polymer membranes, ETS-4 offers sharper size-based separation and thermal stability, though scalability remains a challenge.

Zakeri et al. [65] were able to make a highly N₂-selective ETS-4 membrane by a secondary growth method on porous alpha-alumina tubes. The ETS-4 membranes exhibited an outstanding N₂/CH₄ permselectivity up to 75.2, one of the highest reported values of zeolite-based membranes. The improved selectivity was explained by the accurate molecular sieving effects enabled by the ultramicroporous nature of ETS-4 and the close intergrowth in the membrane layer. The permselectivity exhibited in this work is high, and this shows the importance of diffusivity-based separation in ultramicroporous structures. Although CH₄ has a stronger adsorption affinity to ETS-4, the membrane was kinetically selective to N₂. This means that size exclusion, as opposed to adsorption strength, was the prevailing mechanism, which is a desirable but scarce property of zeolitic membranes in N₂/CH₄ separation applications.

This finding is consistent with other zeolite-based membranes like AlPO-18 and SAPO-34, where N₂ permeance was also improved because of their smaller kinetic diameter. Nevertheless, the capacity of ETS-4 to sustain such a high selectivity, even at low pressure and temperature changes, makes it stand out in the precision of molecular sieving. In contrast to other membranes (e.g., AlPO-18 M-2), which are plagued by structural defects or pressure-related selectivity losses, ETS-4 membranes showed good reproducibility between samples (M1-M3). In addition, the membrane has the unique property of CH₄-preferential adsorption, but N₂-selective permeation, which offers a twofold benefit to adjust the transport behaviour according to the operating conditions, which is not commonly found in N₂-selective materials.

The potential of ETS-4 to provide both high N₂ selectivity and excellent reproducibility in a typical fabrication pathway provides the possibility of investigating titanosilicate-based structures beyond the traditional uses. Specifically, the incorporation of ETS-4 into mixed-matrix or composite membranes may be used to overcome the typical trade-off between permeability and selectivity in N₂/CH₄ separation. The summary of these key insights is presented in Table 9.

The high N₂/CH₄ permselectivity shown by ETS-4 membranes, especially via kinetic sieving in the face of CH₄-preferential adsorption, indicates a largely unexplored route to the design of gas separation membranes. Future research should aim at optimizing the microporous framework of ETS-4 through elemental substitution or ion-exchange to achieve even greater size-exclusion selectivity, with the permeance drawbacks typical of ultramicroporous frameworks being minimized. Also, the hybridization of ETS-4 with polymeric or metal-organic frameworks (MOFs) may produce new MMMs that combine the kinetic selectivity of ETS-4 and enhanced processability and permeability. Considering the temperature and pressure sensitivities, process modelling studies are also suggested to assess the feasible performance of ETS-4 membranes in the industrial conditions of varying impurities, pressure, and long-term use.

5.2.8 Carbon molecular sieve (CMS) membranes

The synthesis of carbon molecular sieve (CMS) membranes based on different polyimide precursors is still an important step towards the advancement of gas separation technology. Three important studies provided complementary information on the impact of pyrolysis conditions, polymer structure and ageing behaviour on membrane performance. Fu et al. [66] performed a study on the transport and sorption characteristics of CMS membranes prepared using a new series of polymer precursors using a pyrolysis procedure that was carefully controlled. The 6FDA-based polyimides produced CMS membranes through a three-step process. Polyimides were first prepared by the polycondensation of 6FDA with diamines (e.g., DAM, DETDA, 1,5-ND) in NMP and subsequently imidized chemically with acetic anhydride and 8-picoline. The powders obtained were dried and redissolved in THF or NMP to prepare dense films, which were subsequently annealed at 210 °C to eliminate the solvent. Lastly, the films were pyrolyzed in argon up to 550 °C with a controlled heating program to produce CMS membranes to be tested in gas separation. TGA demonstrated that three of the polymers were thermally stable at 400 °C, whereas 6FDA/DETDA: DABA(3:2) had early weight loss because of decarboxylation of DABA units, which may facilitate crosslinking. The 6FDA/1,5-ND: ODA(1:1) had the highest decomposition temperature (Td). DSC measurements showed different Tg values, and 6FDA/DETDA: DABA(3:2) showed the most favourable 8 (Td-Tg) to form hollow fibers. Values of D-spacing reduced following pyrolysis, and did not have a direct connection to CMS permeability. Densities of polymer and CMS were in agreement with structural variations, and FFV values provided information about polymer, but not CMS membranes.

Experimental and carbon molecular sieve (CMS) membrane performance measurements of N₂/CH₄ separation (Fig. 4) provide important information on how polymer structure, free volume, and ageing affect the gas transport properties. First, the pure polyimide membranes exhibited different N₂ and CH₄ permeabilities and selectivity. 6FDA: BPDA(1:1)/DETDA had the highest permeabilities (2.625 and 2.550 GPU, respectively) but the lowest selectivity (1.03), which can be explained by its large d-spacing (6.9 A) and high fractional free volume (FFV = 0.182), which favours gas permeation over size discrimination. Conversely, the lowest FFV (0.147) and densest packing, 6FDA/1,5-ND: ODA(1:1), had the highest selectivity (2.00), but very low permeabilities (1.500 and 7.500 GPU). 6FDA/DETDA and 6FDA/DETDA: DABA(3:2) had moderate performance, with a balance between permeability and selectivity. When the membranes were pyrolyzed to make CMS membranes, the permeability increased dramatically, especially in 6FDA/DETDA: DABA(3:2), which had the highest pre-ageing permeabilities (N₂:6.51 GPU; CH₄: 5.42 GPU). Nevertheless, 6FDA/DETDA, the least permeable, was the most selective (N₂/CH₄) of all (1.52 before ageing and 1.65 after one month), owing to ageing-induced densification that constricts micropores and improves size-sieving. This ageing effect had a small effect of enhancing selectivity in all samples by decreasing CH₄ diffusivity relative to N₂. Finally, although CMS membranes enhance flux, 6FDA/DETDA is still the best choice in high-selectivity applications, particularly following ageing, which further demonstrates the significance of structural control and post-treatment in the design of gas separation membranes.

CMS Membrane Gas Separation Performance (Pre- and Post-Ageing)

Full size image

Another study on Matrimid membrane was also carried out by [28], who developed CMS dense film membranes from Matrimid^® 5218 commercial polyimide to separate N₂ from CH₄. The study evaluated the permeation, sorption and diffusion properties of dense CMS films to understand the high selectivity observed in CMS membranes. The study reported that the Matrimid^® 5218 was dissolved in dichloromethane, mixed for 6 h, cast in a Teflon dish, and dried under controlled conditions. The resulting film was cut into discs and pyrolyzed under argon at 550, 675, or 800 °C using a stepwise heating protocol, which were then analyzed after cooling.

The permeation results show that Matrimid polymer dense films become more permeable to N₂ after pyrolysis at 800 °C while preserving high N₂/CH₄ permselectivity. The 800 °C carbon molecular sieve (CMS) dense film derived from Matrimid precursor exhibits 21 times higher N₂ permeability at 50.85 GPU compared to 2.4​​ GPU Matrimid precursor and 6.74 times better N₂/CH₄ permselectivity at 7.69 compared to 1.14. The observed improvement in permeability is primarily due to an increase in the N₂ sorption coefficient (from 3.94 to 81.34 cm³ (STP)/cm³ CMS) and the Langmuir hole-filling capacity. On the other hand, the increased diffusion selectivity (Dₙ/Dₘ) from 3.05 to 15.3 is responsible for the high permselectivity. The temperature-dependent permeation tests between 25 °C and 50 °C revealed that both N₂ and CH₄ permeabilities increased, but permselectivity decreased from 8.27 at 25 °C to 7.57 at 50 °C. This is because CH₄ transport exhibits higher temperature sensitivity than N₂ transport. In general, the experimental findings demonstrate that CMS membranes pyrolyzed at 800 °C deliver superior N₂ permeability alongside high N₂/CH₄ permselectivity than conventional polymer membranes.

While progressing in the effort of exploring various polymeric precursors for CMS membrane production [30], reported a research findings on the investigation of N₂/CH₄ gasses separation on P84 HFCM at different temperatures (298, 323, and 373 K) to further observe the role of temperature dependence on the gas separation performance of carbon membrane by the thermodynamic and activation energy analysis. The P84-copolyimide was prepared and soaked in water, ethanol-treated for solvent exchange, and dried. For further modification, pyrolysis under N₂ involved stabilization at 300 °C, heating to 700 °C at controlled rates, and natural cooling. XRD analysis revealed that HFCMs experienced decreased d-spacing values from 3.69 to 3.96 Å as heating rates increased, which indicated densification of the structure. The structures formed better when reaction rates remained low but higher rates produced imperfect graphite structures. FESEM confirmed dense, defect-free membranes with a retained polymeric structure and thicker layers at higher heating rates.

The permeation performance of the material depended on both heating rate and temperature conditions. As shown in Fig. 5, the N₂ permeability reached its maximum value of 6.40 GPU at 3 °C/min and 298 K, but increased to 1.40 GPU at 373 K, which resulted in a N₂/CH₄ selectivity of 9.09. The CH₄ permeability decreased as heating rates increased because the d-spacing decreased, which demonstrated molecular sieving behaviour. Selectivity was highest at 3 °C/min due to optimized micropore formation. The higher activation energy of N₂ at 13.91 kJ/mol compared to CH₄ at 8.37 kJ/mol led to enhanced selectivity during diffusion. The high performance of these membranes exceeded the Robeson curve, which indicates feasibility for industrial utilization.

Effect of Heating Rate and Permeation Temperature on Gas Permeability and Selectivity of HFCM

Full size image

5.2.9 Comparative CMS membrane performance for N₂/CH₄ separation

A brief comparative overview of the three CMS membrane studies is given in Table 10., highlighting the main differences in precursor materials, maximum N₂ permeability and selectivity, structural effects, thermal behaviour and ageing behaviour. It points out the manner in which every study has maximized various parameters to attain expected trade-offs between the flux and separation performance. Fu et al. [64] showed that high permeability, especially in polymers with large fractional free volumes such as 6FDA/DETDA: DABA, tends to undermine selectivity.

It was inferred from the examined literature that precursor structure and thermal treatment are the key factors that influence the performance of CMS membranes. All the studies, however, differ in their optimization strategy. Ning and Koros concentrated on the influence of high carbonization temperatures, Widyanto et al. investigated the effects of heating rate on pore formation, and Fu et al. paid attention to the molecular-level tuning and ageing effects. Nevertheless, all the studies highlight the possibility of CMS membranes in high-efficiency N₂/CH₄ separation.

Summarily, it is evident from these studies that controlled structural tuning, whether through polymer design, pyrolysis conditions, or post-treatment, consistently leads to improved N₂/CH₄ selectivity through enhanced size-sieving and diffusivity discrimination. Moving forward, future research should target mixed-gas studies under humid or industrially relevant conditions, evaluate mechanical durability, and develop scalable fabrication techniques such as hollow fibre spinning. Hybrid strategies that integrate polymer design with controlled pyrolysis and post-treatment may ultimately bridge the gap between high permeability and long-term selectivity, paving the way for commercial deployment.

5.2.10 Frontier inorganic membrane materials for N₂/CH₄ separation

According to a study carried out by [32], a new class of 2D materials known as MXenes sheets were chosen as the lamellar membrane materials for N₂/CH₄ separation based on double selectivity mechanism. The authors fabricated Cr-MXene membranes through the deposition of MXene colloidal solution containing different Cr cluster concentrations onto Nylon substrates modified with polydopamine. The resulting membranes were dried and activated under vacuum to expose unsaturated Cr sites for enhanced performance. MXene membranes without Cr clusters were also fabricated for comparative purposes. The results of characterization tests validate the synthesis process of Cr-MXene membranes as Cr clusters expand the interlayer spacing and diminish crystal structure order. The XRD peak shift indicates structural expansion. The TEM-EDS analysis shows Cr deposition while XPS results demonstrate fluorine termination which ensures strong bonding. Measurements of cross-sectional images show the membrane thickness expanded from 842 nm to 1080 nm as this enabled the creation of 2D transport channels for better performance.

The MXene membranes lacking Cr clusters had poor N₂ permeance (55 GPU) and moderate selectivity (2.67) due to restricted free space. Introducing Cr clusters increased the d-spacing, which improved gas transmission. Before activation, Cr-MXene membranes showed a trade-off: while N₂ permeance increased dramatically (up to 344 GPU at 5 mg Cr loading), selectivity reduced to 1.98 due to larger pore size, allowing both gases to permeate more freely (Table 11). Cr clusters introduced unsaturated sites following activation that improved both N₂ adsorption and transport selectivity. The best performance was reached with 5 mg activated Cr-MXene which resulted in 381 GPU N₂ permeance and 13.76 selectivity. The activation process enhanced N₂ interactions while maintaining sufficient transport capacity. The excessive Cr-N₂ interactions at 7.5 mg activated Cr loading limited N₂ desorption, which resulted in decreased N₂ permeance to 352 GPU and selectivity to 11.07.

It can be observed from the study that the Cr-MXene membranes exhibit a better permeability-selectivity trade-off, critical to high-throughput gas separation, compared to HOF-21/6FDA-DAM MMMs (Ma et al. [21]), which had a higher selectivity (20.5) but much lower permeance (5.13 GPU). In contrast to MOF-based CH₄-selective membranes (e.g., Ni-MOF-74/SBS or ACU/PVA), which are focused on maximizing methane enrichment with relatively low selectivity (34), Cr-MXene is aimed at energy-efficient N₂ removal, which is a less popular, but industrially important, pathway to natural gas upgrading. Convergence is in the usage of metal sites (Crⁿ⁺, Zr⁴⁺, Ni²⁺) to tune sorption selectivity. Unlike embedded-filler MMMs, Cr-MXene utilizes 2D lamellar stacking and post-synthetic activation, enabling tunable channel structure and reactivity. The most important trade-off is noted between interlayer expansion and selectivity. Overloading Cr increases d-spacing and decreases molecular discrimination, whereas optimal activation provides synergistic increase in permeability and selectivity. This is a new benefit of tunable performance by activation that is not observed in conventional MMMs. The stabilization of structure under pressure and humidity, the incorporation into polymer matrices to form hybrid structures, and the viability of scale-up to real gas mixtures should be the subject of future work. The Cr-MXene system is an encouraging model in the next-generation, N₂-selective membrane design.

5.3 Mixed-Matrix membranes (MMMs)

5.3.1 Metal–Organic frameworks (MOFs)

Metal-Organic Frameworks (MOFs) are types of crystalline porous materials that are constructed using metal ions or clusters connected by organic ligands [67]. They are very versatile and can be used in many applications as their modularity enables them to be controlled in precise pore size, surface area, and chemical functionality. MOFs have become a topic of great interest in recent years in the gas separation processes as a result of their outstanding adsorption capabilities, the tunability of their structure and their ability to selectively transport gases [22, 68]. Specifically, limited work has been reported on the use of MOFs for N₂/CH₄ gas separation; however, their combination with other materials as a mixed matrix has been reported in this regard [20]. A number of studies have been devoted to the design of MOFs with custom pore structure, open metal sites, or ligand functionalization to increase N₂/CH₄ selectivity via membrane-based processes. The purpose of this review is to discuss the current progress in the application of MOFs in the separation of N₂/CH₄ with a focus on experimental results and computational predictions.

In view of this, Bhasker et al. [69] examined the application of MOFs as fillers in systematically selected polymers to enhance the separation characteristics of the resulting composite membranes in the N₂/CH₄ gas separation process. In the present work, MMMs were synthesized based on MOFs, with synthesized CuBTC and PBI polymers. The CuBTC was synthesized through a solvothermal process with BTC and copper acetate in DMF/EtOH/H₂O solution, and then washed and dispersed in DMAc. PBI variants (PBI-BuI and PBI-HFA) were prepared by polycondensation in polyphosphoric acid, purified, and in some cases N-substituted. The PBI solution was mixed with CuBTC suspension, cast and evaporated to produce dense MMMs. In the case of dual-layer asymmetric membranes, a PBI-BuI base layer was cast, and then a mixed solvent system of ZIF-8/PBI-HFA was cast on top of the base layer and phase inverted in water. All the membranes were dried and cut into circular samples to be used in gas permeation tests. Good MOF-polymer compatibility was evidenced by the well-preserved CuBTC structure (WAXD) and even MOF dispersion with no agglomeration or interfacial gaps (SEM, elemental mapping) in the composite membranes. TGA showed that MOFs were thermally stable at 300 °C because of degradation, whereas polymers were stable at 520 °C. Mechanical tests indicated decreased tensile strength at increased MOF loading, but flexibility and modulus were retained in certain variants. The agreement between experimental and theoretical densities also pointed to dense, porosity-free membrane structures.

Figure 6 (permeability) and Fig. 7 (selectivity) are bar charts that show how loading CuBTC and polymer structure influence N₂ and CH₄ gas separation. In all membrane series, N₂ and CH₄ permeabilities were low relative to the lighter gases, as would be expected due to their larger kinetic diameters and less strong interactions with the polymer matrix. Figure 6 indicates that in the CuBTC@PBI-HFA series, both N₂ and CH₄ permeability slightly increases with 30% CuBTC addition. This however, as Fig. 6 shows, resulted in relatively small variations in N₂ and CH₄ selectivity, which varied between 2.27 and 3.00. This implies that although CuBTC increases the overall gas transport due to increased diffusivity or sorption, its impact on selectivity is not high in this system. In the CuBTC@PBI-BuI membranes, on the other hand, the permeability rose more significantly with loading, but selectivity decreased steadily (Fig. 7), falling to 1.69. This implies that the addition of CuBTC caused microstructural defects or interfacial voids that interfered with molecular discrimination, as it was also found in SEM analysis.

In the case of CuBTC@DMPBI-BuI series, a different trend was observed. N₂ Permeability did not vary significantly with filler loading as depicted in Fig. 6, but N₂ and CH₄ selectivity rose slightly between 2.25 and almost 2.88 at 30% loading (Fig. 7). This suggests enhanced dispersion and interaction of the particles at low to moderate loadings, probably because of enhanced compatibility between the methyl-substituted polymer and the CuBTC framework. Conversely, the CuBTC@DBzPBI-BuI composites had the largest baseline permeabilities (Fig. 6), but a steep and steady decline when CuBTC was added. It is noteworthy that N₂/CH₄ selectivity was low all along, and it never went beyond 1.53 (Fig. 7). The decreased permeability and selectivity indicate the pore blockage by the bulky t-butylbenzyl side groups that can impede the gas transport and obscure the inherent properties of the MOF. Hence, the plotted trends highlight the strong dependence of membrane separation performance on both polymer-filler interactions and structural compatibility. Among the studied systems, DMPBI-BuI at 30% CuBTC loading presents the best compromise between permeability and N₂/CH₄ selectivity, while DBzPBI-BuI composites showed limited benefit from CuBTC addition.

Permeability of N₂ and CH₄ gases through CuBTC-incorporated PBI-based composite membranes at varying CuBTC loadings (0–30 wt%)

Full size image

N₂/CH₄ selectivity of CuBTC@PBI composite membranes as a function of CuBTC loading (0–30 wt%)

Full size image

Gu et al. [27] also reported a study on the use of ZIF-8 with carbonized surface ZIF-8@VR, which was then blended with polyacrylic acid (PAA) to fabricate PAA/ZIF-8@VR MMMs for CH₄/N₂ separation. The MMM production involved PDMS and PVA layer deposition onto polysulfone support to create a modified polysulfone (MPSf) membrane with permeable hydrophilic characteristics. The MMMs contained ZIF-8@VR at two different weight percentages (60 and 66.8 wt%) while their thicknesses ranged from 390 to 800 nm. The water dispersion of ZIF-8@VR underwent ultrasonication followed by mixing with PAA before stirring and degassing and storage. The researchers fabricated pure PAA and MPSf membranes alongside their study for comparative analysis. Characterization result confirmed that ZIF-8@VR synthesis was successful while preserving crystallinity alongside surface carbonization. The SEM/TEM images revealed uniform hexagonal shapes that grew in size after the sintering process. The analyses with XRD, FTIR, and Raman spectroscopy proved that both structural integrity and the introduction of defects had occurred. The surface area decreased based on BET analysis results while CH₄ adsorption increased. Acid resistance improved at pH 2.8. SEM/EDS analysis of MMMs demonstrated that ZIF-8@VR particles distributed evenly throughout the material which improved the interaction between the polymers.

The permeation performance results (Fig. 8) demonstrate that PAA/ZIF-8@VR MMMs exhibit improved CH₄ permeance and CH₄/N₂ selectivity compared to MPSf and pure PAA membranes. Increasing ZIF-8@VR loading from 60 to 66.8% significantly enhances CH₄ permeance, reaching a maximum of 2900 GPU at 66.8% loading and 600 nm thickness. The highest CH₄/N₂ selectivity (3.12) is observed for PAA/66.8%-ZIF-8@VR-800, attributed to surface carbonization and enhanced gas adsorption properties.

Comparative Gas Permeance and Selectivity of PAA/ZIF-8@VR-Based Membranes

Full size image

The research shows that PAA/ZIF-8@VR MMMs achieve better CH₄/N₂ selectivity and permeance as a result of the carbonized ZIF-8’s impact on gas transport properties. However, further optimization steps are also required to achieve proper permeability-selectivity balance as aggregation becomes more likely at higher loading levels. Testing the membrane’s durability along with capacity to scale up production and its operational performance using combined real gas streams would enhance practical deployment possibilities. Additionally, the performance of membranes can be improved through investigation of new polymer matrices when combined with post-synthetic modifications.

[64], [29] developed Ni-MOF-74/SBS MMMs with the goal of enhancing CH₄/N₂ selectivity while examining the effect of temperature and Ni-MOF-74 loadings on separation performance. The Ni-MOF-74 employed in this investigation was synthesized using a solvent-thermal technique, and Ni-MOF-74/SBS MMMs (2–20 wt%) were produced using solution casting, with SBS dissolved in an organic solvent and Ni-MOF-74 distributed separately. The solutions were cast onto glass plates and dried, resulting in membranes of a particular thickness. XRD revealed Ni-MOF-74 incorporation into SBS, while FTIR validated typical functional groups. SEM revealed homogeneous dispersion at low loadings, but aggregation above 20 wt%. Nitrogen adsorption demonstrated microporosity (1272 m²/g). TGA found two decomposition stages, consistent with SBS and Ni-MOF-74 degradation, with compositions that matched theoretical values. The study reported that when the Ni-MOF-74 concentration increased, CH₄ and N₂ permeability decreased. However, CH₄/N₂ selectivity remained consistent at around 3.0. At 15 wt% loading, Ni-MOF-74’s inherent affinity with CH₄ and the development of dual gas transport channels resulted in a peak CH₄ permeability of 1.11GPU (116% higher than SBS). However, at 20 wt%, permeability decreased due to particle agglomeration, resulting in reduced free volume. Increased temperature (25–55 °C) improved permeability (CH₄: 0.95 to 1.99 GPU) by increasing polymer chain flexibility and free volume. However, CH₄/N₂ selectivity falls marginally, indicating a trade-off between permeability and selectivity at high temperature. Binary gas studies validated Ni-MOF-74’s capacity to improve CH₄ permeability. Compared to other MOF-based MMMs, Ni-MOF-74/SBS-15 demonstrated superior CH₄ permeability (2.35GPU) and competitive selectivity.

The research indicates that the integration of Ni-MOF-74 into SBS membranes improves CH₄ permeability by enhancing diffusivity; however, excessive loading (> 15 wt%) results in particle agglomeration, which diminishes permeability. Elevated temperature improves permeability while somewhat lowering selectivity, which indicates a trade-off between the two. Future studies should optimize filler dispersion, verify long-term stability, and validate the scalability of Ni-MOF-74 for CH₄/N₂ separation under industrial circumstances.

5.3.2 Comparative studies of MOF MMM membranes in N₂/CH₄ separation application

The recent developments of metal-organic framework (MOF)-based MMMs in the N₂/CH₄ separation indicate that important progress has been made in the design of materials, but also indicate that there are trade-offs between permeability, selectivity, and structural integrity. Critical comparison of major studies provides a more balanced picture of their relative strengths and weaknesses and future prospects. Another common pattern in all the studies reviewed is the improved CH₄ transport by the addition of MOFs, which is mainly attributed to the improved diffusivity and sorption affinity. In the ZIF-8@VR-PAA system of Gu et al., surface carbonization of ZIF-8 increased CH₄ adsorption, driving permeance to 2900 GPU, but with moderate selectivity (3.12). In the same way, Ni-MOF-74/SBS MMMs developed by Wang et al. [64] demonstrated a compromise between moderate selectivity (~ 3.0) and enhanced CH₄ permeance (2.35 GPU) by taking advantage of the inherent affinity of Ni sites to methane. These results highlight a common approach of using MOFs that have selective CH₄ interactions and designed pore structures to enable separation. The other converging knowledge is the importance of dispersion and filler loading. All the studies observed that above a certain concentration (usually 1520 wt%), MOF agglomeration has a detrimental impact on permeability and structural homogeneity, as in [64] and [69]. Therefore, the best filler loadings and good polymer-filler compatibility are still the basis of significant performance improvements.

In spite of these convergences, the studies differ greatly in separation performance trends and design decisions. As an example, Bhasker et al [69]. showed that CuBTC enhanced permeability modestly and only offered modest selectivity enhancement in most PBI variants (max ~ 3.0). In other cases, selectivity was reduced by interfacial voids (e.g. CuBTC@PBI-BuI), and DMPBI-BuI provided the optimal balance at 30% loading, a unique instance of permeability and selectivity increasing simultaneously. Conversely, Gu et al. showed that surface modification (carbonized ZIF-8) can maintain MOF integrity, increase compatibility and adsorption affinity, and thus lead to increased permeance (up to 2900 GPU) without compromising selectivity (3.12). This indicates that a possible way to overcome the common permeability selectivity trade-off is through post-synthetic MOF tuning. Wang et al. [64] emphasized the temperature sensitivity of MMMs. The high temperatures enhanced CH₄ permeability by 109% and reduced selectivity, indicating a trade-off that is determined by the mobility of the polymer chain. This indicates the necessity of high-temperature rigid matrix designs that are thermally stable. These discrepancies show that selectivity improvements are very sensitive to the quality of polymer-MOF interactions, filler dispersion, and operating conditions, whereas permeability improvements can frequently be more easily attained through MOF inclusion alone. The surface-carbonized ZIF-8 by Gu et al. is one of the reviewed work that showed that surface engineering of MOFs can not only enhance gas adsorption but also interfacial compatibility, resulting in membranes that are more selective and permeable than their unmodified counterparts. Bhasker et al [69]. also give an important observation: the structure of the substituent in the polymer is a critical factor. As an example, bulky side groups (DBzPBI-BuI) inhibit performance through blocking transport pathways, whereas methylated structures (DMPBI-BuI) improve both dispersion and selectivity. This highlights the need of co-design of both filler and matrix.

For further research, the following directions seem the most promising: Post-synthetic MOF modifications (e.g., surface carbonization, defect engineering) should be prioritized to increase compatibility and functional selectivity. Interfacial interaction can be optimized, and voids can be removed by the design of smart polymers that are specific to the MOF chemistry, such as flexible chains with functional groups to anchor them. Morphology could be stabilized, and gas separation selectivity could be extended beyond the present limits by hybrid membranes based on dual-layer structures or interfacial crosslinking. Applicability should be guaranteed by more studies reporting real gas mixture performance, membrane ageing/stability, and resistance to industrial contaminants. Materials screening using machine learning could speed up the search for the best polymer-MOF combinations and loading limits of particular gas pairs.

5.3.3 Covalent organic frameworks (COFs)

Covalent organic frameworks (COFs) are crystalline porous materials made of light elements (including B, C, N, and O) by means of strong covalent bonds [25]. They are appealing to gas separation applications due to their highly ordered pore structures, large surface areas, and tuneable chemical functionalities [70, 71]. Nevertheless, it is a major challenge to fabricate defect-free pure COF membranes, particularly on an industrial scale. Consequently, the majority of the research on COF application in N₂/CH₄ separation has been aimed at integrating COFs into polymer matrices to create MMMs, which combine the benefits of the two materials. In this regard, COF-derived MMMs have demonstrated potential to improve selectivity and permeability through the addition of size-sieving effects and gas-specific interactions with the framework. This review is an overview of the latest developments in the use of COF-based MMMs in N₂/CH₄ separation, focusing on experimental results and trends in performance.

Hongwei et al. [72] prepared a bilayer covalent organic framework (COF) membrane through a temperature-swing solvothermal process by successively growing imine-linked COF-LZU1 and azine-linked ACOF-1 on a dual-amino-functionalized Al 2 O 3 disk. COF-LZU1 (~ 0.94 nm pores) was synthesised at room temperature, and then ACOF-1 (~ 1.8 nm pores) was synthesised at 120 °C, leading to an interlaced pore structure with a smaller effective pore size (0.3–0.5 nm), as confirmed by N₂ adsorption-desorption isotherms. SEM, PXRD, and EDX mapping characterisation showed a defect-free membrane of ~ 1 μm thickness and high crystallinity and good intergrowth. The surface area was 386 m²/g as determined by BET analysis. The layered stacking was verified by energy-dispersive X-ray spectroscopy, which did not show any boundaries between the COF layers. The membrane had a high molecular sieving behaviour in terms of gas separation. It had an ideal selectivity of 105.0 and mixed gas selectivity of 100.2 at 298 K and 1 bar, especially in N₂/CH₄ separation. The bilayer structure was much more selective and mechanically stable compared to single-layer COF membranes. The membrane exhibited separation performance over a wide range of temperatures (up to 393 K) and extended operation (> 100 h), which indicated high thermal and operational stability because of the strong covalent structure.

Also, the computational study of [13]considered the gas separation potential of a range of nanoporous materials including covalent organic frameworks (COFs) with special attention to the CH₄ and N₂ separation. Although the original study reported the results in terms of CH₄/N₂ selectivity, re-analysis of the results in terms of N₂/CH₄ separation offers more insight into how COF membranes perform in applications that require nitrogen rejection of methane-rich streams. The findings indicated that COF membranes had N₂/CH₄ selectivity of about 0.13 to 0.91, the opposite of at 1 bar pressure.

This means that, under most conditions, COF membranes selectively permeate CH₄ over N₂, indicating a solubility- or adsorption-controlled separation process, with the higher condensability and interaction of methane with the pore walls overcoming the diffusivity advantage normally enjoyed by smaller N₂ molecules. Nevertheless, even with this CH₄ affinity, the paper has identified a specific COF membrane, NPN-3, which has shown the best adsorption performance score (APS) of all COFs, 40.2 mol/kg, which is a very strong uptake potential that can be optimized to be selective in future designs. COFs also showed a similar CH₄-selective trend at high pressure (10 bar), with recalculated N₂/CH₄ selectivities remaining below 1, and confirming that none of the tested COF membranes showed a preference towards nitrogen over methane under the simulated conditions. Although the selectivity of N₂ over CH₄ is suboptimal, the study highlights a significant discovery: COFs have a high capacity to absorb methane and can be tuned to have high porosity, and with structural adjustments (e.g., pore window sizes, functional group integration, or flexibility of the framework), they can be re-targeted to preferentially permeate N₂. This has great potential in the development of MMM, where COFs with suitably designed interfaces and molecular sieving properties could be used to supplement polymer matrices, which already have some N₂ affinity.

5.3.4 COF membranes for N₂/CH₄ separation: experimental and computational perspectives

Experimental and computational studies of the application of covalent organic frameworks (COFs) to N₂/CH₄ separation provide convergent evidence of their potential usefulness, especially because of their adjustable porosity and large surface area. One of the most important overlaps is that COFs, both in pure membrane and in MMMs form, have considerable structural benefits that can be utilized in gas separation. As an example, the bilayer COF membrane developed by Hongwei et al. had an extraordinary selectivity (ideal selectivity of 105.0 and mixed gas selectivity of 100.2), indicating that properly designed COFs can provide high molecular sieving effects that are beneficial in nitrogen rejection. This confirms the overall consensus that structural design (e.g., layered stacking, pore narrowing) is key to COF membrane performance.

However, there exists a gap in the behaviour of separation in different studies. Although the experimental COF membrane was N₂-selective, Gulbalkan et al. computational study consistently revealed CH₄-preferential permeation over a broad range of COF structures. Simulated N₂/CH₄ selectivities recalculated (0.13 0.91) show that the majority of the modeled COFs preferred to transport methane, which implies that in idealized conditions and common COF topologies, CH₄ adsorption would be dominant because of its greater condensability. This points to a trade-off between permeability and selectivity, and in practice, performance can vary with fabrication quality and structural modification.

One new observation of these joint studies is the discovery of the NPN-3 COF membrane, which, although it exhibited CH₄ selectivity in simulation, had a high adsorption performance score, suggesting that it may be a powerful candidate to tune N₂ selectivity by functionalization or integration in selective polymers. Moreover, the bilayer membrane has been shown to be successful in the experiment, which indicates that a combination of various COFs to tune the effective pore size is a potential approach to selective molecular sieving. However, it is worth noting that most of the insights, particularly those that are based on simulations, rely on idealized assumptions and are not experimentally tested. Trends in computational selectivity should thus be viewed with caution, and future studies should focus on experimental synthesis and testing of best-performing candidates to verify their feasibility in practice.

The next step in the development of COF-based membranes in N₂/CH₄ separation is to address fabrication issues to obtain scalable, defect-free membranes. The next step should focus on optimizing pore size and functionality to prefer N₂ diffusivity, and the synthesis of high-performing computational COFs such as NPN-3 into the real world. The combination of COFs with N₂-selective polymers in MMMs is a viable route, although the interfacial compatibility should be enhanced. In general, experimental confirmation, stability over time, and scalable processing will be essential to transition between concept and industrial use.

5.3.5 Exploring frontier approaches in MMMs to N₂/CH4 separation

The recent breakthroughs in material design and membrane engineering have resulted in a new generation of MMMs that are specifically designed to perform the challenging task of N₂/CH₄ separation. These new developments, which can be summed up under the title Frontier Approaches in MMMs to N₂/CH4 Separation, present a variety of innovations that have potential industrial applications.

For instance, Ma et al. [21] reported the synthesis of MMMs of HOF-21 (a hydrogen-bonded organic framework) nanofillers in a 6FDA-DAM polyimide matrix to separate N₂/CH₄. The current study aimed to enhance the N₂-permeable membranes, which can significantly save the energy cost of recompression in the natural gas purification process. The N₂ diffusion was enhanced by the HOF-21 material that has a pore size of approximately 3.6 Å due to its favourable molecular sieving effect. Moreover, because HOF-21 possessed nearly identical adsorption affinities to N₂ and CH₄, it suppressed CH₄ solubility, and thus increased N₂/CH₄ solubility selectivity. The MMM exhibited a N₂ permeance of 5.13 GPU and N₂/CH₄ selectivity of 20.5 at 50/50 vol% % feed gas at 0.2 MPa and 298.15 K at an ideal HOF-21 loading of 7 wt%. These values have respective increases of 67% in N₂ permeance and 54% in selectivity over the neat 6FDA-DAM membrane, which shows that the synergistic effect of HOF-21 is significant on both the diffusivity and the solubility selectivity. This paper demonstrates that HOF-based MMMs can be used to make high-performance N₂-selective membranes to purify CH₄ in an energy-efficient way.

Furthermore, Liang et al. [73] recently developed a new type of MMM that was specifically designed to enrich low-concentration CH₄ to increase the separation of CH₄/N₂ mixtures in unconventional energy applications. The membranes were made by a simple dip-coating method, with a porous polytetrafluoroethylene (PTFE) membrane serving as a mechanical support. The selective layer was a mixture of styrenebutadiene styrene (SBS) block polymer and nickel-based metal-organic framework (Ni-MOF-74), which was prepared by hydrothermal reaction of 2,5-dihydroxyterephthalic acid and nickel nitrate hexahydrate. SEM confirmed the uniform morphology of the Ni-MOF-74 particles with sizes of 500–800 nm, and XRD analysis confirmed that the particles maintained their crystalline structure, which means that they remained structurally intact within the membrane matrix. When this filler was incorporated into the SBS polymer matrix, which was reinforced by PTFE, a mechanically reinforced MMM with high interfacial compatibility was obtained. It is noteworthy that the membrane containing 20 wt% Ni-MOF-74 loading (PM20) exhibited a high mechanical strength of 37.7 MPa, which was much higher than that of its unreinforced counterparts, indicating the efficiency of the PTFE-supported structure. The PM20 membrane had a CH₄ permeability of 6.90 × 10⁻⁸ GPU and CH₄/N₂ selectivity of 4.18 in terms of gas separation performance. This implies that methane is preferentially permeated over nitrogen, and thus it is appropriate in CH₄ enrichment processes. The findings indicate that the large surface area and microporosity of Ni-MOF-74, flexibility and processability of the SBS polymer played a role in improved permeability and moderate selectivity. Not only did this work show a feasible method of fabricating scalable membranes, but it also highlighted the potential of reinforced MMMs in the recovery of methane in dilute gas streams, which presents an exciting pathway to energy-efficient gas purification technologies.

Additionally, Yu et al. [74] designed a new type of MMMs to enhance CH₄/N₂ separation by using amorphous UIO-66-NH₂ with carbonized structure (ACU) as the inorganic filler. The parent UIO-66-NH₂ metal-organic framework was calcined under high-vacuum conditions to cause carbonization and form abundant Zr⁴⁺ active sites to synthesize the ACU. The SEM analysis revealed that this treatment caused structural changes to coarsened, amorphous particles of 300–380 nm in size, compared to uniform cubic crystals. The obtained ACU nanoparticles had better chemical stability and alkaline resistance due to the creation of a strong carbon matrix, but were still porous enough and had a high CH₄ affinity. These structural properties led to high CH₄/N₂ selectivity based on Ideal Adsorbed Solution Theory (IAST). Notably, the ACU fillers showed high compatibility with polyvinyl alcohol (PVA) polymer matrix, in which the membranes were formed. To fabricate the final membrane matrix, the ACU nanoparticles were evenly dispersed in PVA to create defect-free MMMs, and a polysulfone (PSf) support was used to give mechanical stability. The PVA/ACU-120 A MMM with 80 wt% of the optimized variant of ACU-120 A was found to be the best-performing membrane among the membranes tested, with an outstanding CH₄ permeance of 5584 GPU and CH₄/N₂ selectivity of 3.43. This is a considerable improvement over the common polymer-based membranes and shows that the addition of thermally stabilized amorphous MOF-derived fillers can be used to enhance the gas transport properties without reducing the integrity of the membrane. The paper also points out the possibility of carbonized MOF derivatives as a novel type of filler in high-performance gas separation MMMs, especially in methane enrichment.

To further address the challenging separation of CH₄ from N₂ due to their closely related physicochemical properties, a novel mixed matrix membrane was fabricated by Wang et al. [75] through the incorporation of a Hofmann-type metal-organic framework (MOF), CoNi-DABCO, into a PDMS polymer matrix. The CoNi-DABCO MOF features a narrow pore size and a framework rich in oppositely positioned open metal sites, which impart a strong binding affinity toward CH₄ molecules. This structural design promotes preferential CH₄ adsorption and facilitates its selective permeation across the membrane. The membrane fabrication involved dispersing varying loadings of CoNi-DABCO particles into the PDMS matrix, with the optimal performance observed at a loading of 20 wt%. The resulting MMMs were characterized for their structural and transport properties. The incorporation of the CoNi-DABCO filler significantly enhanced the adsorption capacity for CH₄ without compromising the membrane’s mechanical integrity or processing compatibility. At 20 wt% loading, the membrane achieved a CH₄ permeability of 1285 Barrer, coupled with a CH₄/N₂ mixed-gas selectivity of 3.7, demonstrating the dual benefit of enhanced permeability and reasonable selectivity. Moreover, the membranes showed excellent pressure resistance up to 10 bar and stable long-term performance over 30 days, highlighting their potential for practical gas separation applications. This study underscores the effectiveness of Hofmann-type MOFs as functional fillers in developing high-performance MMMs for CH₄ enrichment and N₂ removal.

A new approach was also introduced in a recent study by Gu et al. [27] to enhance the separation performance of membranes in CH₄/N₂ separation by developing a highly engineered MMM system. The researchers employed a surface-carbonized and stiffened zeolitic imidazolate framework-8 (CSZ), which was obtained through high-vacuum-resistance calcination (HVRC). The surface modification strategy increased the structural rigidity of the material and produced Zn-rich active sites in the carbonized structure. These characteristics played a major role in the selective adsorption of CH₄ and enhanced thermal and chemical stability; thus, CSZ is a potential filler material in gas separation. The CSZ nanoparticles were dispersed uniformly in a PVAm matrix. The synergistic effect of CSZ and PVAm was observed, where the two materials enhanced the preferential adsorption of CH₄ compared to N₂, which was caused by the strong binding interactions on the Zn-rich sites and the amine functional groups. The interaction was confirmed by gas adsorption isotherms and density functional theory (DFT) calculations that showed increased CH₄ uptake relative to N₂. The resultant membranes showed high scalability and mechanical strength, and the authors were able to prepare large-area MMMs of up to 1120 cm² and module designs of about 2000 cm². Remarkably, the membrane attained an excellent CH₄ permeance of 7600 GPU and a CH₄/N₂ selectivity of 4.35 at 85 wt% CSZ loading. These values are a huge improvement on most traditional polymeric membranes, particularly when it comes to methane enrichment of nitrogen-containing gas mixtures. Besides showing the promise of carbonized ZIF-based fillers such as CSZ, the work also shows that high-performance, scalable MMMs can be produced to be used in industrial gas separation processes.

5.3.6 Frontier strategies and emerging paradigms in MOF-Based, MOF-Derived, and HOF filled MMMs for N₂/CH₄ separation

Despite the convergence in aims of the reviewed studies on new materials development for N₂/CH₄ separation, the methodology employed in terms of the type of fillers, membrane structure, and performance focus differs greatly. Ni-MOF-74/SBS and CoNi-DABCO/PDMS are designed to achieve moderate CH₄ selectivity and permeability by increasing adsorption, ACU/PVA and CSZ/PVAm focus on very high permeance, aiming at the recovery of dilute CH₄, with moderate selectivity. By contrast, HOF-21/6FDA-DAM aims at N₂ enrichment, which is comparatively scarce and a reverse separation rationale. Moreover, the fabrication techniques vary, including hydrothermal synthesis of fillers, high-vacuum calcination (HVRC), dip-coating and layered support membranes, which demonstrates a diverse environment of processing developments.

Nevertheless, one of the trends observed in the studied literature is the replacement of traditional polymer-based membranes with filler-based MMMs, which are aimed at breaking the permeability-selectivity trade-off in N₂/CH₄ separation. The majority of these works use metal-organic frameworks (MOFs), especially ZIFs, Hofmann-type MOFs, and MOF derivatives, because they have tunable pore structure, large surface area, and the possibility of selective gas adsorption. The exploitation of CH₄-selective pathways is also converging, with most of the studies focusing on this direction, with one (HOF-21) making a significant shift towards N₂-selective separation to save energy. Also, nearly all studies indicate high levels of filler-polymer interfacial compatibility, a key parameter to reduce non-selective voids and stable membrane performance.

In the course of the review, a number of cross-cutting insights are revealed through the comparative analysis. The two materials, ACU and CSZ, obtained through the thermal processing of MOFs, have enhanced polymer compatibility, stability, and selectivity. This implies that carbonized MOF derivatives may form a novel filler frontier. In addition, the use of PTFE supports of SBS-based membranes by Liang et al. and the large-area CSZ/PVAm modules by Gu et al. indicate a trend toward scalable manufacturing, which is an aspect that is not usually considered when developing MMM. Moreover, the near-equal adsorption of CH₄ and N₂ and enhanced N₂ solubility selectivity of HOF-21 indicate that the solubility selectivity modulation, and not diffusivity, is a feasible pathway, particularly in N₂-selective separations.

Based on these insights, a number of future research directions are apparent. A dual-mode filler design offers a chance to create fillers that combine the synergy of diffusivity selectivity, and adsorption affinity, particularly through the incorporation of multi-functional MOFs or hybrid porous carbons. Selective membranes to separate N₂ are not yet well developed, and since most existing membranes prefer CH₄, the HOF-21 study represents an exciting pathway to energy-efficient N₂-selective MMMs. Stability in mixed-gas and high-pressure environments should be addressed further; although certain studies mention stability (e.g., CoNi-DABCO demonstrating 30-day performance), long-term ageing, cycling, and real gas stream performance are underreported and crucial to commercialization. Another frontier is membrane module integration and scale-up; Gu et al. were the first to demonstrate module-scale membranes, and more work should be done to scale MMMs beyond lab-scale films to membrane modules, such as hollow fiber configurations. Finally, AI-aided MMM design and optimization, via machine learning tools to screen filler/polymer pairs, predict permeability/selectivity performance, and optimize filler loadings, is a new but underdeveloped frontier.

This comparative analysis (Table 12) emphasizes that while major progress has been made in the development of MMMs for N₂/CH₄ separation, particularly via MOF-based approaches, a new development of strategies, emphasizing N₂ selectivity, scalable processing, dual-function fillers, and real-world robustness, is needed to fully unlock their industrial potential.

The literature reviewed shows that MMMs especially those that include MOFs, zeolites, or carbonized frameworks within engineered polymer matrices, hold the most promise of energy-efficient and scalable N₂/CH₄ separations. Nevertheless, the shift between the laboratory-scale prototypes to the industrial-scale modules requires a critical assessment of technical viability and economic feasibility.

6.1 Material cost and availability

Most of the best fillers employed, including HOF-21 (Ma et al. [22]), Ni-MOF-74 (Liang et al. [20]), carbonized ZIFs (Gu et al. [23]), and CuBTC (Bhasker et al. [69]), are based on relatively inexpensive precursors, such as metal nitrates and carboxylic acids, and employ scalable synthesis methods (e.g., hydrothermal, calcination). Nonetheless, solvent (e.g. DMF, DMAc, polyphosphoric acid) and post-synthetic modifications can add cost and safety issues to scale-up. Therefore, it will be necessary to optimise green synthesis procedures and low-toxicity solvents.

6.2 Fabrication scalability and module design

Some of the studies showed scalable strategies in membrane fabrication. As an example, Gu et al. [23] were able to fabricate MMMs with module areas of over 2000cm², which is a good sign of the ability to manufacture at scale. Liang et al. [20] used dip-coating on PTFE supports, implying that it could be compatible with roll-to-roll or continuous casting processes, which are already in industry. Conversely, MMMs based on fragile polymers (e.g., PBI in Bhasker et al. [69]) or multilayer structures may necessitate more complicated process control, which can add capital costs unless simplified through extrusion or lamination methods.

6.3 Performance vs. Process cost Trade-Off

In terms of the separation process. As demonstrated by Ma et al. [22] and Montes Luna et al. [65], N₂-selective membranes have the potential to save a lot of recompression energy in CH₄ recovery processes, as the inert N₂ is removed at the front-end of the membrane, and no downstream pressurization is required. On the other hand, CH₄-selective membranes such as those of Yu et al. [69], Wang et al. [25], and Gu et al. [23] are more appropriate in biogas upgrading where enrichment of CH 4 is important. Therefore, the membrane design selected will determine the overall energy demand and compression cost of the plant depending on the gas source (natural gas vs. biogas). Membranes with moderate selectivity (3.5) and high permeance (> 1000 GPU), such as in Gu et al. and Yu et al. are particularly appealing to compact module design and smaller footprint, which equates to lower capital and operating costs. For instance, assuming a target of 1,000 Nm³/h, a membrane with 4× higher permeance could reduce module surface area (and cost) by up to 75%. This supports the case for high-permeance MMMs in cost-sensitive or small-footprint applications.

6.4 Mechanical and thermal stability

Scalability is also determined by the long-term stability of the membranes in terms of pressure and temperature. Investigations of Gu et al. [23] and Wang et al. [25] emphasized that MMMs could perform at 10 bar and 55 °C, respectively. Moreover, mechanical flexibility and pressure resistance, essential to industrial applications, were achieved by the use of polymers, such as SBS, PVAm, and PAA, in combination with strong fillers, such as carbonized MOFs.

6.5 Fouling resistance and maintenance costs

While not extensively explored in the reviewed studies, the addition of hydrophilic or functionalized fillers (e.g., surface-carbonized ZIFs in Gu et al. [23]) and smooth and defect-free MMMs (e.g., Yu et al. [69]) could provide some fouling resistance. Nevertheless, durability testing at industrial scale in multi-component or sour gas mixtures remains unavailable and this needs to be done to estimate the frequency of membrane replacement, a major cost driver.

6.6 Lifecycle and sustainability considerations

Membranes such as those reported by Liang et al. [20] and Yu et al. [69] rely on thermally or chemically modified MOFs, which introduce recyclability challenges. In the meantime, the application of stable polymers and MOFs based on common metals (e.g., Ni, Zn, Zr), and relatively low calcination temperatures makes MMMs more sustainable than conventional amine or cryogenic systems. Nevertheless, the end-of-life management, membrane disposal and reusability should be considered in the techno-economic modelling.

The recent surge in innovative membrane materials, particularly the integration of metal-organic frameworks (MOFs), hydrogen-bonded organic frameworks (HOFs), carbonized MOF derivatives, and zeolites into robust polymer matrices, has laid a promising foundation for advancing membrane-based N₂/CH₄ separations. Despite these advancements, several technical, operational, and economic challenges must be addressed to bridge the gap between laboratory success and industrial adoption. This section outlines specific future research directions to guide continued innovation in membrane-based nitrogen/methane separation, as recommended by the peer reviewer.

7.1 Toward Process-Specific membrane design

The studies reviewed illustrate divergent membrane targets, either CH₄-selective (e.g., Ni-MOF-74/SBS, ZIF-8@VR/PAA, ACU/PVA, CSZ/PVAm) or N₂-selective (e.g., HOF-21/6FDA-DAM, clinoptilolite/PBI). Future work should prioritize application-specific design: N₂-selective membranes for natural gas upgrading to reduce recompression costs, and CH₄-selective membranes for biogas and landfill gas enrichment. Hybrid systems that combine both membrane types sequentially may also be explored to improve overall process efficiency.

7.2 Enhancing Permeability–Selectivity Trade-off

Although a few MMMs have shown outstanding CH₄ permeance (e.g., > 5000 GPU in ACU/PVA, > 7000 GPU in CSZ/PVAm), selectivity often remain modest. There is a clear need to exceed the Robeson upper bound by improving both diffusivity and sorption selectivity. Potential solutions are post-synthetic MOF functionalization to add selective adsorption sites, the use of dual-filler or hierarchical filler structures, and the creation of mixed-linker MOFs or flexible frameworks that deform under pressure.

7.3 Stability, scalability, and compatibility

The long-term chemical, thermal, and mechanical stability of MMMs under industrial gas streams (containing CO₂, H₂S, water vapour, etc.) remains underexplored. While Gu et al. and Liang et al. demonstrate some scale-up viability, systematic upscaling studies involving long-term permeation under varying pressures, module fabrication and integration into pilot-scale systems, and mechanical durability testing under cyclic stress are essential before industrial adoption.

7.4 Interfacial engineering and defect control

Many studies (e.g., Bhasker et al., Wang et al.) noted challenges with filler-polymer interfacial compatibility, leading to non-ideal morphologies and performance drops at higher loadings. Future work should leverage on surface functionalization of fillers (e.g., amino, carboxyl, carbonyl groups), polymer side-chain engineering to improve affinity, and the incorporation of interfacial compatibilizers to eliminate non-selective voids.

7.5 Green manufacturing and lifecycle sustainability

A growing emphasis on sustainability demands that future membrane development focus on green synthesis of fillers by low-energy or solvent-free processes, non-toxic and biodegradable polymers or supports, and end-of-life membrane recycling and regeneration. These will be decisive in techno-economic modelling and in the alignment of membrane development with global sustainability goals.

7.6 Integrated Techno-Economic and process simulation studies

Although material performance is significant, validated cost-performance models needs to be implemented in practice. Comparative techno-economic evaluations of membrane systems and existing cryogenic or PSA processes, process simulation (e.g., Aspen Plus) using real membrane data, and lifecycle cost analysis (CAPEX, OPEX, maintenance) to determine the most promising membrane systems to implement are urgently needed.

This review has examined recent developments in the membrane-based technologies of N₂/CH₄ separation, with a particular focus on emerging materials including PIMs, zeolites, MOFs, HOFs, and MMMs. The overlapping trends in the studies indicate that the optimized filler dispersion, interfacial compatibility, and the tailored pore structures have a great impact on the gas selectivity and permeance. Although membranes with HOF-21 or Ca²⁺-clinoptilolite showed better N₂-selectivity, others like Ni-MOF-74/SBS or CSZ/PVAm were more CH₄-selective, which indicates the different design approaches depending on the end use. MMMs that use thermally stable MOFs or zeolites were the most promising of the reviewed studies in terms of scalable deployment, and they have both mechanical integrity and high separation performance. Nevertheless, the majority of studies are confined to laboratory-scale tests that lack long-term validation, and the performance data are usually hard to compare due to a lack of uniformity in reporting standards.

From a techno-economic perspective, the feasibility of industrial application will depend on scalable membrane fabrication, durability under real gas conditions, and consistent performance under pressure and contaminant exposure. Future studies should focus on standardization of testing procedures, long-term stability studies and integrated system-level evaluations to bridge the gap between material innovation and process application. In summary, as filler design and membrane engineering continue to improve, especially via the MMM platform, high-performance and scalable membranes to separate N₂/CH₄ are becoming more feasible, a promising route to energy-efficient purification of natural gas.