Chapter 1: Introduction to Functional Membranes

With the increase in population growth, global warming, environmental degradation, and resource shortage, there is a huge requirement for reducing environmental emissions; this has prompted increased attention to the development of efficient separation technologies for the “zero discharge” requirement and “green” production. The development of the sustainable separation technologies is one way to reduce emissions and achieve “zero discharge” and lower environmental pollution. Functional membranes form a separation technology that uses a porous or nonporous material barrier to selectively separate molecules or ions from an initial feed stream into a purified permeate stream and a rejected retentate stream. Membrane separation technology is a promising strategy for separating a wide range of fluids, including gas, oil, and water.¹  It offers attractive advantages in mass separation processes or chemical conversion processes due to the relatively low energy requirements, facile operation, small footprint, scalable capacities, high separation efficiency, and no requirement of additional chemicals.² 

A broad classification system based on the fluid phase of the feed and permeate components distinguishes each membrane separation. Gas separation membranes have a gaseous feed and permeate. Similarly, liquid separation membranes have a liquid feed and permeate. However, pervaporation membranes or membrane distillation processes have a liquid feed and gaseous permeate. Liquid membranes include aqueous separation membranes, oil–water separation membranes, and organic solvent separation membranes. Aqueous separation membranes have an aqueous water feed and permeate. Oil–water separation membranes have a mixed organic solvent and aqueous water feed and an organic solvent or aqueous permeate. Organic solvent separation membranes have an organic solvent feed and permeate (see Figure 1.1a). The membrane structure (pore size) and driving force for mass transport (e.g., pressure, concentration, etc.) can also be used to further distinguish gas, liquid, and pervaporation separation membranes. For example, depending on the pore size, membrane classifications are reverse/forward osmosis (RO/FO), electrodialysis/ion exchange (ED), nanofiltration (NF), ultrafiltration (UF), and microfiltration (MF) (see Figure 1.1b).^3–5 

As one of the most energy-efficient separation technologies, membrane technology can play a significant role in sustainable development.^6,7  Although many traditional separation technologies, including evaporation, precipitation, crystallization, salting out, solvent extraction, and ion exchange have been used in industry, their accompanying high costs and energy use, and low separation efficiency limit their sustainability.^8–10  Membrane separation processes utilize a fine porous or nonporous structure that offers attractive advantages such as relatively low energy requirements, facile operation, small footprint, scalable capacities, and high separation efficiency, in mass separation or chemical conversion.^11,12  In particular, membranes with finely porous and nonporous structures, specific separation performance, eco-friendly potential, and high energy efficiency lead to scaled-up applications. Thus, membrane separation processes offer attractive advantages in mass separation or chemical conversion.^13,14  Therefore, membrane separation technologies have gained importance as an affordable and feasible alternative to conventional separation technologies.

Nowadays, membrane technology is well developed and widely used in numerous industrial processes, including gas separation, aqueous separation, pervaporation, oil–water separation, and organic solvent separation.^15–19  Also, membrane separation technology is currently being applied in novel applications, such as specialized biomedical diagnostics and therapy or energy conversion processes (e.g., fuel cells and alternative batteries). These applications have been extensively investigated at lab scale and are in development for commercialization.^20,21  Each application requires different membrane properties as shown in Table 1.1.²²  Polymers and inorganic materials have been used in membrane fabrication to yield specific membrane properties, and easily scalable manufacturing processes such as phase inversion, interfacial polymerization, and coating have been developed for these materials. In effect, the number of industrial applications for membrane technology has increased (e.g., water reuse projects, see Figure 1.2), but stricter government regulations and the creation of the “zero discharge” industry requirement have prompted the need for more selective and permeable membrane performances.²³  Following the new requirements of high selectivity and “zero discharge” industry, state-of-the-art of materials and manufacturing processes have been proposed and synthesized in recent years.^24–26  Therefore, many more developments in materials science will be required for making advanced membranes and investigating their membrane processes from a long-term perspective.

Conventional membrane separation processes face several challenges and limitations. In recent years, most researches deal with very relevant aspects of such development, either from the point of view of a special membrane application or process, or with a focus on certain classes of materials or fabrication methods. Considering the state-of-the-art and on-going efforts on the membrane technology development, the challenges and limitations of membrane materials and preparation, membrane performances for essential applications, and membrane sustainable manufacturing processes will be discussed and used as the guidance for the future of membrane technology.

Originally, membranes were fabricated using a small number of established materials, and some were not explicitly intended for membrane separation technology. Most traditional membranes have a porous asymmetric polymeric structure and are prepared by phase separation of a polymeric casting solution. The two types of phase separations for membrane fabrication are non-solvent-induced phase separation (NIPS) and temperature-induced phase separation (TIPS). Spinodal decomposition or nucleation and growth are the mechanisms that lead to pore formation. During phase separation, the segregation of the polymer-deficient domains and the gelation–solidification of the polymer-concentrated domains affect the pore sizes and uniformity.²² 

To increase the sustainability of traditional membrane separation technology, novel renewable materials should be explored for membrane fabrication, and membrane fabrication processes should minimize their use and waste of hazardous chemicals. Currently, commercial membranes’ production utilizes organic solvents for dissolving the polymers used in phase separation casting solutions or the monomers used for interfacial polymerization. Environmental regulations aim to reduce the use of these solvents, and more environmentally benign solvents are required for membrane fabrication. Studies on replacing conventional solvents with less harmful ones are currently in investigation.

While the sustainability of membrane fabrication is hampered primarily by the solvents used, the membrane material and materials used for module fabrication also affect the sustainability of membrane separation technology. A cradle-to-grave approach is necessary, where the membrane process and the fabrication of the membranes and modules, as well as their fate after use, should be taken into consideration. Membranes with a prolonged lifetime and a stable process performance can reduce wastes and enable module reuse, while in other applications, biodegradable and disposable membrane modules may offer a more sustainable solution.

Recently, numerous membrane processes have been designed and built for wastewater treatment, target ion-selective separation, surface water treatment, desalination, sterile filtration, chloro-alkali electrolysis, food and beverage processing, hydrogen separation, nitrogen enrichment, and natural gas sweetening. Besides, the construction of new membrane processes plants is expected to increase in the coming decades for other proposed separations, such as specialized biomedical diagnostics and therapy or energy conversion processes (e.g., fuel cells and alternative batteries). However, there are many critical challenges in membrane separations to tackle, which will require improved or novel membranes.

Currently, most membranes face challenges related to energy consumption, fouling (including surfactant fouling, organic fouling, inorganic fouling, and biofouling), etc., as shown in Table 1.2. For some membranes, there are specific challenges (e.g., target ion selectivity of ion-exchange membranes and target gas separation by the gas separation membranes), which are summarized in Table 1.3. Notably, commercial membranes can significantly benefit from improved intrinsic separation properties such as selectivity and permeability. There is a natural trade-off between these two properties, most expressed for gas separation membranes as the ideal selectivity versus the membrane material's permeability in the “Robeson Upper Bound”. Similar relationships have been shown for gas separation, NF, and RO, where there is overall decreasing permeance with increasing solute rejection.

A further challenge for mature membrane applications is the need for sustainable application processes. Sustainability can be seen in a holistic way, including process scalability, energy saving, and economics. New membrane separation process technologies are needed to minimize energy consumption and obtain a high separation efficiency. Recently, numerous membrane separation plants have been built in many regions, and the construction of new membrane separation plants is expected to increase in the coming decades.²²  For better separation and purification, improved efficiencies and conservation are vital to meet growing demand and improving availability. Combining different membrane separation technologies or integration with different membrane separation processes may be increasingly applied for future separation processes.^21,22 

Gas separation (GS) membranes separate gaseous mixtures in a pressure-driven process. GS membranes typically have an anisotropic single-phase or composite structure with a thin nonporous selective layer, ranging from 0.1 to 0.5 μm, on top of thicker microporous supports. GS membranes typically have hollow fiber and spiral wound membrane modules. Hollow fiber membranes are fine capillaries with diameters that range from 50 to 500 μm. Spiral wound membranes are flat sheets that are rolled tightly. Most commercial GS membranes have hollow fiber configurations because they are much cheaper to produce. Compared to conventional gas separation techniques such as pressure swing absorption and cryogenic distillation, the main advantages of GS membranes are their low environmental footprint and elimination of costly gas-to-liquid phase changes in the gaseous mixture.^27–29 

Most commercial GS membranes are made of polymeric materials and have a dense selective layer on top of a thicker, porous support. This dense layer has no continuous passages or pores through the membrane for molecular transport. Instead, the diffusing molecules pass through the selective layer in transient gaps formed by thermally agitated relocation of the polymer matrix. Therefore, the dominant transport mechanism for gaseous separations is solution–diffusion. The three steps in the solution–diffusion are the diffusing molecules’ sorption into the selective layer, migration through the selective layer, and desorption at the membrane's low-pressure side (see Figure 1.3). Solution–diffusion depends on the diffusing molecule's solubility and mobility and favors the smallest and most condensable molecule.

The performance of GS membranes depends on the permeability of the target component (x) in the gaseous mixture through the membrane (P_x) and the selectivity of the target component (x) against other components (y) in the gaseous mixture (P_x/P_y). There is a well-known trade-off between selectivity and permeability in GS membranes, known as the Robeson upper bound (see Figure 1.4). Established in 1991, Robeson's upper bound explains that polymers with higher gas sorption have larger segmental mobility, leading to higher permeability but lower selectivity. This upper bound was quantified in plots of log(P_x/P_y) versus log(P_x) for O₂/N₂, H₂/N₂, He/N₂, H_2/CH₄, CO₂/CH₄, and He/H₂ gaseous mixtures. Robeson's upper bound was updated in 2008 to account for the state-of-the-art materials for gas separations, and computational analysis and simulations continue to explore new materials that push this boundary. Commonly used materials for GS membranes include polymers with high glass transition temperatures, such as glassy polymers and rubbery polymers. These polymers or molecules have higher chain rigidity and larger free fractional volume, which helps increase selectivity and permeability. Some examples are cellulose acetates, polyphenylene oxide, polytetrafluoroethylene, polycarbonates, polyimides, and polysulfone. Current research for novel GS membrane materials concentrates on hyper rigid backbone polymers, polymers of “intrinsic micro-porosity” (PIMs), crosslinking methods for polymers, and porous additives in composite membranes or mixed matrix membranes such as metal–organic frameworks, zeolites, and covalent organic frameworks. Research on these novel materials is promising for extending the current applications of GS membranes. Overall, GS membranes continue to attract interest for their highly sustainable processes, and with active research, they may become the leading technology for gas separations.

Pervaporation (PV) membranes separate liquid mixtures into vapor components in a chemically driven process. PV membranes also adopt an anisotropic single-phase or composite structure with a thin nonporous selective layer on top of a thicker microporous support. The dominant configuration for PV membranes is the plate and frame module, where a group of flat sheet membranes, porous supports, and spacers are layered together between two plates. PV processes are ideal for mixtures that are not easily separable by conventional distillation, extraction, and sorption methods. In particular, PV membranes are suitable for separating mixtures with azeotropes, close-boiling points, isomers, and thermally sensitive compounds. Moreover, hybrid processes that integrate PV membranes with other liquid separation technology are emerging in industry. Other process variants include thermal PV, perstraction, osmotic distillation, saturated vapor permeation, and pressure-driven permeation.^30,31  Membrane distillation is a very similar process that separates liquid mixtures into gaseous components by a temperature gradient and has applications for aqueous liquid desalination.

PV membranes are unique because there is a low energy phase change from the liquid feed to the vapor permeate. Unlike distillation, only the feed mixture's target component (typically less than 10 wt%) consumes the latent heat as it diffuses through the membrane. There are three steps in the PV membrane process: solution, diffusion, and evaporation. First, the PV membranes absorb the target component from the liquid mixture, diffuse it through the membrane's swollen surface, and release it as low-pressure vapors due to the latent heat consumption. Then, the low-pressure vapor is condensed and collected or discharged in subsequent steps.

The driving force for mass transport is a chemical potential gradient created by applying a vacuum pump or inert purge to the PV membrane's permeate side. The vacuum pump is the more prevalent method. However, the inert gas method is ideal for applications that discharge the permeate. Moreover, the heat for the phase change of the diffusing component comes from the feed solution, leading to an additional heat transfer and temperature gradient. Mass transport also depends on feed composition and concentration, feed pressure, permeate pressure, temperature, and film thickness.

Separation by PV membranes is dependent on the membrane materials, the physical structure of the membrane, physiochemical properties of feed mixtures, and permeate–membrane interactions. Polymeric materials, including glassy polymers, rubbery polymers, and ionic polymers, are commonly used. Glassy polymers are ideal for water-selective membranes used in solvent dehydration, such as hydrophilic poly(vinyl alcohol) (PVA) and poly(acrylic acid). Cellulose polymers, chitosan, alginate-based polymers, and aromatic polyimide materials are also other options. Ionic polymers with ionic groups neutralized by counter ions show a high affinity for water and are used for dehydration membranes. In most cases, these hydrophilic polymeric materials used in dehydration applications are crosslinked or otherwise modified (e.g., blending) to increase stability and selectivity. However, rubbery polymers such as poly(dimethylsiloxane) are more suitable for organic separation from water. Moreover, both rubbery and glassy polymers are used in organic mixture separation.

Non-polymeric materials such as robust inorganic zeolite membranes and microporous hydrophilic amorphous silica membranes are useful for PV processes. Zeolite membranes have uniform and molecular-sized pores and have been successfully commercialized for dehydration membranes.³²  The majority of inorganic membranes are used as support membranes crosslinked to a dense polymeric layer. Other emerging materials are amorphous per-fluoropolymers, carbon molecular sieves, graphene and graphene oxide, and PIMs. The latter three materials have seen more success as filler in mixed matrix membranes.³³  Overall, continuous research on innovative materials and the fundamental science of PV processes will expand this exciting technology's industrial applications.

Aqueous separation membranes separate aqueous solutions in a process driven by a pressure gradient (MF, UF, NF, and RO), chemical (dialysis and FO), temperature gradient (MD), or electrical field gradient (ED). MF, UF, dialysis, and NF membranes have an asymmetric single-component or multi-component composite cross-sectional structure with a thin porous selective layer on top of a thicker microporous support. RO, FO, and ED membranes utilize an asymmetric single-component or multi-component composite cross-sectional structure with a thin nonporous selective layer on top of a thicker microporous support. Specialty ED or dialysis membranes include ion-exchange membranes, which have charged functional groups. MD processes employ a superhydrophobic microporous asymmetric membrane structure with pore sizes ranging from 1 nm to 1 µm.³⁴ 

Owing to the variation in structure and materials, aqueous separation membranes follow two main mass transport mechanisms: pore flow and solution–diffusion. The pore flow mechanism is a simple filtration process that separates molecules by their size; this is a size-sieving effect. The loose or porous membrane structure allows the preferential transport of smaller components in the feed through the pores and rejects the larger components. The solution–diffusion transport mechanism for aqueous separations is the same as for gas separation membranes, except that the feed and permeate phases are different. Dialysis, MD, MF, and UF membranes adhere to the pore flow theory; RO, FO, and ED membranes keep the solution–diffusion mechanism. NF membranes are a combination of both mechanisms. Overall, aqueous separation membranes can effectively remove particles in an aqueous mixture ranging from 10⁻⁴ to 10² μm, including ions, proteins, viruses, bacteria, colloids, and pollens, as shown in Figure 1.4.

The overarching characteristics of practical aqueous separation membranes are high flux for aqueous solvent, high selectivity, low membrane fouling, chemical and mechanical stability, and low cost. Polymers are the most cost-effective material for processing and application, so they dominate the membrane market. Generally, aqueous separations do not require specialty materials due to their more ambient settings and lower thermal and chemical demands.

Therefore, the most common polymers for RO and NF membranes are polyamide thin-film composite membranes and cellulose acetate. Typical UF membrane materials include CA, polyacrylonitrile (PAN), poly(ether imides), aromatic polyamides, polysulfone (PS), poly(ether sulfone) (PES), poly(vinylidene fluoride) (PVDF), and poly(vinyl pyrrolidone). Typical MF membranes use PVDF, PS, polyamides (PA), and polyethylene (PE). The main preparation methods for these membranes are the phase inversion process for single component membranes and interfacial polymerization for thin-film composite membranes. The configuration mode for aqueous separation membranes also varies by the pore size. MF aqueous membranes use a dead-end filtration mode, and usually, UF membranes follow a cross-flow mode. Most RO and NF membranes have spiral wound modules and some hollow fiber configurations.

While advances in aqueous separation membranes have slowed down compared to other membrane fields, novel materials and fabrication methods for improved selectivity and permeability still emerge. For traditional polymeric materials, track etching is a novel fabrication process that irritates polymeric films with charged particles and then passes them through an etch solution to form uniform and regular pores, albeit in a low volume. Stretching is another fabrication method that extrudes a crystalline polymer film, up to 300%, at a temperature close to its melting point and then rapidly cools it. The stretching process is repeated in the perpendicular direction to produce slit-like pores about 20 to 250 nm wide. Polymer modification by coating, grafting or crosslinking, and spin-coating are other methods to functionalize conventional economical polymers.

Novel polymers include self-assembly of isoporous block copolymers, PIMs, and biometric membrane materials. Block copolymers are highly ordered macromolecules formed by the self-assembly of block polymeric species when placed in a selective solvent. Block copolymer membranes have narrow pore size distributions, high porosities, and sharp molecular weight cutoff (MWCO). PIMs are an advantageous material with a high free volume content and resistance to swelling. Biomimetic membranes exhibit high selectivity and permeability by incorporating biological or synthetic materials such as aquaporins and carbon nanotubes into their structure to mimic biological cell membrane construction.

2D materials benefit from atom-sized thickness, controllable pore size between layers, and mechanical strength. Graphene or graphene oxide (GO) has seen rapid developments in recent years into aqueous separation membranes for its electronic properties, high specific surface area, and high mechanical strength. Other 2D materials include self-supporting nanoparticle layers of transition metal oxides, transition metal dichalcogenides, silicates, and clays. These latter materials face significant challenges in scaling up due to poor control in dispersibility and poor mechanical strength. Nanoparticles such as MOFs, COFs, and porous organic frameworks may also be added into polymeric matrices to form additional pore channels as mixed matrix membranes. Lastly, liquid crystalline materials and polymers are also emerging membrane materials with high flux due to the homeotropic alignment of polymer chains in the direction of molecular transport.

The main limitation in this field is fouling caused by the deposition of particulates, colloids, macromolecules, or microbes onto the membrane surface and the formation of a cake layer. Surface and internal fouling can occur in porous membranes, while only surface fouling occurs on dense membranes. To avoid this issue, membranes are modified to increase their surface hydrophilicity. Increasing the membrane surface's hydrophilicity deters a foulant cake layer formation by creating an aqueous barrier between the foulants and the membrane surface. Moreover, these membranes, particularly RO and NF, suffer from swelling due to the contact with a liquid feed phase and undergo significant degradation during their lifetime. However, unless a novel material or processing technique is cost-competitive with conventional materials, membranes will continue to thrive with the existing technology and need to be occasionally replaced.³⁴ 

The term “oil” refers to animal and vegetable oils, fatty acids, petroleum hydrocarbons, surfactants, phenolic compounds, naphthenic acids, etc.³⁵  Oil is a significant environmental pollutant expelled by wastewater in industries, including steel, aluminum, food, textile, leather, petrochemical, and metal finishing. The oil takes the form of free, dispersed, or emulsified droplets characterized by their size. Free oil has droplet sizes larger than 150 μm, dispersed oil droplet sizes range from 20–150 μm, and emulsified oil droplet sizes are less than 20 μm. Conventional methods such as gravity separation and skimming are more efficient for free oil, while unstable oil–water emulsions can be chemically or mechanically dispersed and treated by the gravity method. However, these conventional methods suffer from several shortcomings such as higher cost, use of toxic compounds, large installation space, and the generation of secondary pollutants.³⁶ 

Oil–water separation membranes are particularly attractive for the highly surfactant stabilized oil/water emulsions with droplets smaller than 20 μm. Oil–water separation membranes separate oil–water emulsions in a pressure-driven process. Oil–water separation membranes typically employ MF and UF membrane structures. Therefore, most oil–water membranes have an asymmetric single-component or multi-component composite cross-sectional structure with a thin porous selective layer on top of a thicker microporous support.

Oil–water separation may be carried out with MF, UF, NF, or RO membranes. UF is the most effective membrane process for oil–water separation due to its nominal MWCO, which refers to 90% rejection of the membrane for the smallest solute molecular weight. For UF, the MWCO ranges between 100 000 and 200 000 Daltons (Da), resulting in a high rejection for stable emulsified oil droplets. Microfiltration membranes have a higher flux but are more likely to have a lower oil separation efficiency. In effect, MF and UF are the best oil–water separation membrane choices; they follow the pore flow theory for mass transport.

The main challenge in this field is membrane fouling due to surfactant adsorption and pore-clogging by the droplets.³⁷  There have been significant research efforts for understanding the mechanism of membrane fouling and developing anti-fouling strategies. Therefore, a wide range of materials with hydrophobic/oleophobic, hydrophilic/oleophobic, or amphiphilic materials have been explored.³⁸ 

Polymers are the most prevalent material used in oil–water membrane separations. PVDF, PS, PES, PAN, and CA are used to prepare MF/UF oil–water membranes.³⁶  To avoid fouling issues, polymers with hydrophilic properties such as polyvinylpyrrolidone (PVP) and polyethylene glycol (PEG) are added to membranes by blending or chemical and physical surface modification. Super wetting and anti-wetting surfaces are emerging approaches for oil–water separation membranes. These materials selectively absorb oil or water while altogether rejecting the other.³⁹  Ceramic filtration membranes made of zirconia, alumina, titania, and silica have excellent performance in separating oil wastewater treatments.³⁷  Overall, oil–water separation membranes form a sustainable technology for difficult oily wastewater separations.

Organic solvent separation processes are carried out mainly with nanofiltration membranes (OSN). OSN membranes reject particles and dissolved molecules smaller than 2 nm with an asymmetric single-component or multi-component composite cross-sectional structure with a thin porous selective layer on top of a thicker microporous support. At the industrial scale, hollow fiber, plate and frame, and spiral wound membrane modules have been used for membrane configurations. OSN membranes depend on molecular interactions between the solute and solvent, solvent and membrane, and the solute and solvent's molecular size. Molecular diffusion in OSN membranes is limited by the size sieving effect and the physicochemical interactions between the membrane and feed. OSN membranes are characterized by the pore-theory and solution–diffusion transport mechanisms and serve as an intermediate process that also exploits the solute's different physiochemical properties, including size, charge, and hydrophilicity/hydrophobicity in separation.

OSN membranes are prepared from organic or inorganic materials or from the combinations of both. Polymeric materials are cheaper but suffer from aging and compaction. The benefits of ceramic membranes include compaction resistance, swelling and leaching resistance, lower feed pressures, and easy cleaning. However, ceramic membranes are much more brittle and more challenging to scale than polymeric membranes. Some ceramic materials used include silicium carbide and Zr⁻, Ti, and Al⁻ oxides. Phase inversion (ISA) and thin-film composite processing methods are the two main ways to process OSN membranes. In many cases, crosslinking and other post-processing procedures are employed, such as wet or dry annealing, drying by solvent exchange, and conditioning treatments to improve the stability and selectivity of OSN membranes.

Common materials used in OSN ISA membranes include (i) PAN, and PI, for NF or UF supports; (ii) polyaniline (PANI), polybenzimidazole (PBI), PSF, or sulfonated poly(ether ether ketone) (SPEEK), for NF applications; and (iii) poly(ether ether ketone) (PEEK), and polypropylene (PP), for UF support applications. Similarly, OSN TFC membranes utilize solvent-resistant polymers and inorganic materials for support membranes, including asymmetric PSF, poly (ether sulfone), PAN, poly (vinylidene fluoride), PP, PI, and PBI. Monomers and polymer used to prepare TFC membranes by interfacial polymerization or coating include diamines, acyl chlorides, and poly(ethylene imine) for IP; and polydimethylsiloxane (PDMS) for coatings. Mixed matrix membranes with nanotubes, zeolites, clay, and fullerenes have been applied to OSN applications.

This technology's main limitation lies in the membrane stability in a variety of organic solvents and obtaining a high and reproducible membrane permeability and selectivity for solvent molecules in the 200–1000 g mol⁻¹ range.⁴⁰  OSN membrane development focuses mainly on the (i) robustness of the membrane materials in organic solvents, in harsh and basic conditions, and during module preparation; (ii) the robustness of the membrane materials toward swelling and leaching; (iii) increasing the multi-functionality of the membranes; (iv) increasing the membrane selectivity for difficult molecular separation such as molecules of similar size; (v) modeling tools to measure transport; and (vi) capital investment for the installation of new OSN plants.^40,41  The separation of organic solvent mixtures remains one of the most promising and growing applications for membrane technology and will undoubtedly continue to grow in the upcoming years.

Some major industrial applications of gas separation membranes include hydrogen recovery, nitrogen- and oxygen-enriched air production, acid gas removal from natural gas and syngas, and air and natural gas dehydration. Table 1.4 shows the main industrial applications developed for gas membranes.

Hydrogen recovery helps supply the increased demand for hydrogen gas in refining petroleum, treating metals, producing fertilizer, processing food, and fuel cell technologies. GS membranes selectively recover hydrogen from ammonia purge gases, syngas mixtures containing CO and CH_4, and petrochemical refinery gas streams. Hydrogen recovery from ammonia purge gases is an ideal application for GS membranes because H₂ is much smaller than N₂, yielding high permeability and selectivity. Also, the purge gas lacks hydrocarbon vapors that may cause fouling or plasticization.

In another application, petroleum refinery processes use hydrogen gases to purify oils and products in hydrotreating, hydrocracking, and hydrodesulfurization processes. Government regulations such as the Clean Air Act and heavier crude qualities require recovery of hydrogen gas, and commercial gas separation membranes provide a simple recovery method (see Figure 1.5). In the GS membrane processes, hydrogen permeates preferentially through the membrane, producing a purified hydrogen “permeate” stream and a hydrocarbon enriched “residue” stream. The hydrocarbon-enriched “residue” is pressurized and recovered for use as liquid fuel or further separated to extract more high-value components.

Nitrogen gas has many safety-related applications (e.g., during shutdowns in compressors, pipelines, and reactors; preventing fires and explosions; and avoiding equipment degradation). GS membranes are an excellent option for producing nitrogen-enriched air and small scale-nitrogen production with relatively low purities. In pretreatment nitrogen recovery, agglomerate filters remove particles, water, and oil from compressed air. Then, the filtered air is pressurized and fed into the gas separation membrane separator. The faster gases (O₂, CO₂, water vapor) permeate first and are discharged to the atmosphere, while the slower nitrogen gas is retained and recovered from the retentate feed. Also, GS membranes produce oxygen-enriched air used in chemical, medical, and food packaging applications. Another industrial application of oxygen-enriched air is furnace and burner injection to increase flame temperature, reduce nitrogen volume, and lower energy consumption.

Natural gas requires the removal of acid gases and contaminants to reduce impurities to acceptable levels. Acid gases are mixtures of natural gas containing significant amounts of H₂S, CO₂, or other gases. Some contaminants in untreated natural gas include oil mist, glycol, methanol, early operating drilling fluids, and ultra-fine iron sulfide particles (formed by hydrogen sulfide reaction with iron pipes). Natural gas is typically purified through an amine gas treatment process; however, membrane systems are very permeable to H₂S and CO₂ and can use the wellhead gas pressure as the driving force for separation. Then, byproduct H₂S is sometimes converted to elemental sulfur to generate valuable sulfuric acid.

Likewise, syngas, a synthetic gas produced by the partial oxidation of methane or oligomerization of methane to produce ethylene, must be purified before its use. Syngas is a fuel-gas mixture consisting mainly of H₂, CO, and CO₂. It is emitted when hydrocarbons (such as coal) are gasified and can produce electricity. Syngas is usually converted into CO₂ and H₂ at high temperatures and pressure by the water gas shift, and GS membranes remove the CO₂.

Dehumidification of air and drying of natural gas are essential applications of GS membranes. For example, Permea (Cactus Membrane Air Dryer) has commercialized a highly efficient membrane system for air dehumidification. Their GS membranes remove water from the compressed air in water vapor, while nitrogen, argon, and oxygen continue through the hollow fiber membranes into the separator. A small amount of slower gas passes through the fibers and purges the water vapor through the separator. Then, the leftover liquid is removed via a coalescing filter upstream of the membrane. The Cactus Membrane works by lowering the dew point, the atmospheric pressure below which water droplets begin to condense and dew can form. As the humidity increases, the separator supplies drier air to lower the dew point.²⁸ 

The three main applications of PV membranes are (i) organic solvent dehydration (e.g., alcohols, ethers, esters, acids), (ii) organic compound removal from aqueous solvents (e.g., removal of volatile organic compounds, recovery of aroma, and biofuels from fermentation broth), and (iii) anhydrous organic mixtures separation (e.g., methyl tert-butyl ether (MTBE)/methanol, dimethyl carbonate (DMC)/methanol). Two commercially developed PV membrane applications are the (iv) dehydration of alcohols and solvents and (v) organic compound removal from water.³⁰ 

Organic solvents such as alcohols, ethers, acids, and ketones have been successfully dehydrated through pervaporation processes. In this application, a solvent mixture contains a low concentration of water, and hydrophilic PV membranes are employed to preferentially allow the passage of this water to produce a water permeate and dehydrated solvent retentate. Membrane materials use dipole–dipole interactions, hydrogen bonding, and ion–dipole interactions to increase water attraction and flux through the membrane. Hydrophobic membranes may also be employed when there is a large amount of water with a small amount of solvent to separate.

PV membranes are advantageous for the dehydration of azeotropes (ethanol and 4 wt% water mixture; isopropanol and 12 wt% water mixture) and other alcohol–water mixtures. Traditional separation methods require the addition of cyclohexane before the distillation process. This additional molecule can create impurities, rendering the alcohol product unsuitable for high concentration purity applications such as pharmaceutical production. PV membranes eliminate the need for added chemicals, and this technology has seen successful operations in pilot plants and industrial applications.⁴⁴ 

PV membranes can recover aromatic compounds from aqueous solutions and remove volatile organic compounds from wastewater. PV membrane processes are advantageous compared to traditional adsorption technology because they can handle high concentrations of organics in water and are a continuous and facile operation. Additionally, the low-temperature operation preserves valuable organics by avoiding thermal degradation. In this application, hydrophobic or organophilic membranes allow the preferential passage of organic solvents while retaining the water retentate. However, there is a limited number of hydrophobic materials available for this application, and PDMS is still the hallmark material for organic recovery by pervaporation membranes.⁴⁵ 

Organic–organic separations by pervaporation membranes focus on separating organic mixtures with very low liquid-to-vapor (LVE) curves. The two industrial applications are methanol and methyl tert-butyl ether, and benzene and cyclohexane separations. Traditional methods such as distillation and extraction are less suitable for separating benzene and cyclohexane solutions that have a volatility difference of only 0.6 °C. PV membranes avoid this limitation by using the adsorption selectivity between the double bond of benzene and the membrane's polar groups. Benzene forms double bonds with polar groups of polymer membranes, and polymers with polar groups can facilitate benzene permeation through the membrane. Organic–organic solvent separations using PV membranes are still in development but have many more potential applications.⁴⁶ 

Aqueous separation membranes are the most commercialized functional membranes. The market for aqueous separation membranes is diverse, ranging from desalination, wastewater treatment, microelectronics, micro-dialysis, and food and beverage production.

Aqueous separation membranes form an important technology for wastewater treatment. Depending on the wastewater composition and intended use, different pressure-driven membrane processes are applied. For example, RO aqueous membranes retain all contaminants and molecules such as bacteria and monovalent ions in clean and potable water production. NF membranes separate pigments, sulfates, divalent cations, lactose, sucrose, and sodium chloride from wastewater streams. MF membranes remove particles from waste streams, such as the concentration of valuable proteins, and bacteria and colloid removal. UF membranes have water treatment applications in concentration, fractionation, or removal of oils, organics, and microplastics in industrial filtration. MF, UF, and NF are commonly used as pretreatment steps for RO to alleviate fouling and improve flux stability (see Figure 1.6a).⁴⁷ 

Unlike pressure-driven membrane processes, FO follows a natural osmosis process in which water molecules are drawn from one solution to another through a semi-permeable membrane (see Figure 1.6b). In this case, a highly concentrated draw solution is used to provide a concentration gradient to draw water molecules from the feed solution. This gradient provides the osmotic pressure difference required to drive water molecules from the feed solution to the drawing solution. This process has been applied to many different wastewater sources ranging from domestic, sewage, and industrial use.

Membrane desalination technology is critical for meeting the worldwide demand for clean drinking water. Desalination processes use NF/RO or ED membranes to separate dissolved ions from water. ED processes use an electrical potential to achieve ion migration from dilute solutions to concentrated solutions through ion-exchange membranes, as shown in Figure 1.6d.⁴⁸  When a current is applied on an ED membrane stack, cations migrate to the cathode and anions migrate to the anode, respectively. The cations are then retained by the anion-exchange membrane (AEM). Similarly, the anions are retained by the cation-exchange membrane (CEM). ED desalination provides high water recovery without the use of chemicals. ED's ability to remove ionic and non-ionic components has qualified it to process municipal wastewater, brackish water, industrial wastewater, and even chemical and food industries. However, it is not suitable for high-salinity wastewater treatment because increased ion content increases ED energy consumption and operation costs.

Membrane distillation is another method for desalination. In this technique, water vapor is transported across a hydrophobic microporous membrane based on a vapor pressure gradient across the membrane. As shown in Figure 1.6c, this thermally driven process mainly facilitates the separation of feed solutions with high water content. MD can be used to purify aqueous solutions by evaporating water vapor from the feed solution into the permeate channel while concentrating salts, solids, and non-volatile solutes.

Oil–water separation membranes form an essential technology to remove oil and grease droplets from industrial wastewater and produced water from petroleum industry processes. Some other industries that have oily wastewater as a byproduct include the food and beverage industry that uses water in washing, rinsing, mixing, and pasteurization process as well as metal industries that use water as a coolant and flushing fluid.³⁵ 

Produced water is high-salinity wastewater produced as a byproduct in oil and gas extraction and is currently the most significant source of waste with 250 million barrels per day of produced water compared to 80 million barrels of oil per day.⁴⁹  Produced water contains a mixture of inorganic and organic compounds, and conventional methods have difficulty in removing the small, emulsified oil droplets. Owing to space constraints, compact physical and chemical treatment plants are preferred for off-shore production, and the application of membrane technology in upstream and downstream processes in the petroleum industry has been used to overcome this challenge. Oil–water separation membranes in combination with RO and NF membranes to reduce produced water salinity is a major application of this type of membrane. Furthermore, hybrid oil–water separation membrane and MD processes also show promise in the treatment of oily wastewater.

Organic solvent filtration technology has been widely used in the pharmaceutical industry; base chemicals industries, including petrochemical and basic inorganic production; consumer chemicals industries; and specialty chemicals, including paints, inks, and dyes and pigments. These fields are newer than well-established gas separation and aqueous separation applications fields, and there is still a lot of potential development and expansion.

Organic solvent-resistant membranes can be used in the pharmaceutical industry to isolate and concentrate antibiotics and pharmaceutical intermediates and for peptide synthesis. OSN membrane separation can protect the drug structure, achieve high purity, avoid drug deactivation in conventional separation techniques, and improve product quality. OSN can also be used to recover and reuse organic solvents used in pharmaceutical synthesis and solvation, and it has been estimated that OSN uses 25 times less energy per liter of recovered solvents compared to distillation.⁴¹ 

Currently, the waste stream purification processes take up to 90% of manufacturing costs in chemical companies, creating a significant market for lower cost and more sustainable separation processes. For example, the petrochemical industry can use OSN membranes in their chilled solvent recovery from lube oil production. Moreover, natural extractions from herbs and other plants provide a source of valuable consumer-ready oils, vitamins, antioxidants, flavonoids, terpenoids, carotenoids, and minerals. These natural extractions have applications in health and wellbeing, and organic solvent separation membranes have been used in the extraction and concentration of these compounds and the recovery of solvent used in extraction processes.

Increasing demands on dye purity and dye effluent discharge standards in the textile industry have led to a growing interest in separating dyes and inorganic salts from crude dyes, printing, and dyeing effluents. Membrane separation offers a wide range of applications and the advantage of achieving dye purification (see Figure 1.7). The large amount of impurities mixed with the crude dyes can lead to poor dyeing results, low chromaticity, and large color differences, which greatly affect the quality of dyed fabrics. NF membranes can capture dyes through the synergistic effect of size screening and electrostatic repulsion.

Yan Zhao would like to acknowledge the support provided by the China Scholarship Council (CSC) of the Ministry of Education, P. R. China (CSC No. 201708330281).