Stimuli-responsive membranes—mechanisms, materials and future directions

Abstract

The development of stimuli-responsive membranes, often stated as smart membranes, has garnered increasing attention in recent years owing to their potential in various industrial separation processes and their ability to mimic natural biological systems. Materials suitable for such applications can dynamically adjust their physical and chemical properties, reacting to external stimuli such as temperature, pH, light, or magnetic/electric fields, thereby enabling precise control over membrane microstructure and the dynamic transport of molecules. This review offers an in-depth examination of the chemistry and responsive mechanisms of different stimuli-responsive materials, along with their integration into membrane matrices to optimize performance. Furthermore, it presents a comparative analysis of various types of stimuli-responses, illustrated with pertinent examples. Ultimately, this review highlights the outstanding challenges and future strategies for advancing smart membranes. With ongoing progress in chemistry and materials science, the development of a selective and efficient nanofiltration membrane platform is anticipated to yield significant benefits to advanced separation processes, offering more efficient and integrated technological solutions.

Introduction

Membrane-based technologies stand as a powerful solution for removing trace amounts of emerging contaminants (ECs), including pharmaceutical, organic pollutants, nanomaterials, and personal care products, addressing the challenges posed by rapid industrialization and urbanization through innovative and determined approaches¹. These processes offer advantages such as scalability, high selectivity, a compact design, and cost efficiency. Membranes can be engineered to specifically target pollutants based on factors like size, charge, and chemical characteristics, achieving high removal efficiencies for various ECs. Furthermore, membrane processes are economical, requiring minimal chemical use, reducing the generation of harmful byproducts and simplifying the treatment process. Common membrane types used in wastewater treatment include reverse osmosis (RO), microfiltration (MF), nanofiltration (NF), ultrafiltration (UF), and membrane bioreactors (MBRs), all of which have effectively removed ECs from domestic and industrial wastewater².

Stimuli‑responsive materials have begun to redefine membrane science by embedding dynamic “on‑demand” behavior into otherwise static nanofiltration architectures. When molecules such as thermo, pH‑, photo‑, magnetic‑, or redox‑responsive polymers are grafted onto a porous scaffold, the resulting hybrid can reversibly reshape its pore geometry, alter surface charge, or switch interfacial wettability whenever an external cue is applied^3,4. This capacity to dial flux and selectivity up or down in real time supplies a level of process control that conventional membranes—whose permeability is fixed at the moment of fabrication cannot match.

That tunability is now being harnessed across virtually every separation landscape. In water remediation, CO₂ bubbling, magnetic fields or near‑infrared light trigger instantaneous wettability flips or pore shrinkage, collapsing oil‑in‑water‑in‑oil and water‑in‑oil‑in‑water emulsions in a single pass and initiating self‑cleaning cycles that arrest organic or biofouling^5,6,7,8. At the biomedical scale, thermoresponsive or pH‑responsive networks open and close nanopores with sub‑micron precision, enabling pulsatile drug release, size‑selective protein harvesting, and dynamically stiffening tissue scaffolds⁹. Photochromic or gas‑binding motifs integrated into industrial membranes provide light‑gated O₂/N₂ partitioning and CO₂/H₂ capture without the energy penalties of cryogenic or pressure‑swing methods, offering a sustainable alternative for chemical and energy plants^5,9. Meanwhile, food‑processing and pharmaceutical lines exploit charge‑switchable composites to purify peptides, vitamins and active pharmaceutical ingredients at kilogram‑per‑hour scales with previously unattainable yields, slashing solvent use and waste generation¹⁰.

Across these disparate arenas, the common thread is a membrane that cycles through multiple permeability states without structural fatigue, positioning stimuli‑responsive platforms as indispensable tools for cleaner, more precise, and resource‑efficient separations in the decades ahead¹¹.

Several reviews have explored advancements in stimuli-responsive membranes, focusing on their fabrication processes and applications^{3,12,13,14,15,16,17}. A recent review emphasizes the development of smart membranes that adjust their performance in response to stimuli and discusses their applications across fields like environmental protection, medicine, and energy¹⁵. Additionally, there is a paper addressing advances in stimuli-responsive membranes for nanofiltration, covering their preparation methods, classifications, and the challenges and future directions in this area¹⁶. Another review highlights membranes that adapt their properties to improve separation performance through various stimuli such as ions, light, pH, temperature, and electric or magnetic fields. It also details preparation techniques like blending, casting, polymerization, self-assembly, and electrospinning¹⁷. Despite substantial progress in understanding external triggers, there remains a notable gap in knowledge regarding the molecular chemistry underlying these materials, including the structure of responsive materials and interactions and their influence on membrane performance for building smarter and more efficient nanofiltration/ultrafiltration membrane platforms.

In this review, we concentrated on how the chemistry of stimuli-responsive materials can be modified to achieve specific functional changes in membrane characteristics such as permeability, wettability, and flux. This review also explores the molecular mechanisms by which the chemical structure and composition of materials affect their ability to respond to external stimuli. By analyzing existing fabrication techniques and material integration strategies, we address the critical challenges in durability and scalability, providing insights into future directions for exploiting these smart materials in industrial and biomedical fields. The review sections will primarily examine materials exhibiting responsiveness to light, electricity, pH, and temperature, as well as the membranes they facilitate. In addition, we will investigate the existing research gaps in membrane development strategies by comparing various categories of stimuli-responsive materials based on their distinct chemistries and their integration into membrane matrices. Ultimately, we will propose strategies for enhancing the scalability of responsive membranes across diverse fields (Fig. 1).

Fig. 1: Focus of this review.

This review addresses four key aspects of stimuli-responsive membranes: the materials used, response chemistry and mechanisms, the regulation of controllable properties, a comparative analysis of various responsive materials, and research gaps and strategies for future membrane development.

Light-responsive materials

Light-responsive molecules are materials that undergo reversible structural changes when exposed to specific wavelengths of light. These molecules can switch between two or more isomeric forms with distinct physical and chemical properties. The structural changes are typically triggered by the absorption of light energy, causing molecular rearrangements such as isomerization, cyclization, or bond breaking/forming. Their exceptional capacity for reversible transformations in response to light exposure is a phenomenon that holds significant implications for the development of advanced materials. Table 1 compares different light-responsive materials (azobenzenes, spiropyran, anthracene, diarylethene, and stilbenes) in terms of their structure, activation energies, isomerization types, and wavelengths. A detailed comparative description of their respective chemistries and light-dependent responsive behavior mechanisms is given in supporting information.

Table 1 Comparison of different light-responsive molecules and their structure, isomerization wavelengths, isomerization type

In this section, the use of light-responsive materials to create smart membranes will be discussed, and how the responsive behavior of different materials affects the selectivity, efficiency, wettability, and permeability of membranes. Mechanisms of light activation of various molecules are presented in Figs. S1 and S2.

Light-responsive molecules enabled membranes

Light-responsive molecules have revolutionized the design of functional membranes by introducing the ability to modulate membrane properties dynamically and reversibly through light stimuli^18,19,20. These specialized membranes, activated by light, exhibit tunable permeability, selectivity, and surface properties, enabling precise control over molecular transport and separation processes. By incorporating light-responsive molecules such as azobenzenes, spiropyrans, and diarylethenes, these membranes offer unique advantages in fields requiring on-demand regulation, such as chemical separations and environmental applications. The versatility and non-invasive nature of light as a stimulus make these membranes particularly attractive for advanced material applications^21,22.

The membrane pore size is a critical factor for achieving optimal separation performance in membranes, as it directly influences the efficiency and selectivity of the separation process. Azobenzene is particularly advantageous due to its high responsiveness to light stimuli. Light-responsive covalent organic network (CON) membranes were developed through the interfacial polymerization of azobenzene-4,4′-dicarbonyl dichloride (AB) and 1,4,7,10-tetraazacyclododecane (cyclen) under optimized conditions (Table 2, entry 1)²³. These membranes are utilized in the selective separation and removal of dye by controlling the trans-cis isomerization under UV/Visible light exposure. The trans membranes resulted in an “on” state with larger membrane pore apertures of around 1.06 nm, while UV light exposure induced cis membranes, shrinking the pore sizes to around 0.75 nm in the “off” state (Fig. 2). Due to their sizes, the trans-CON membranes demonstrated permeance of 22.6 L/m²/h/bar and rejected indigo carmine with 94.7% efficiency. Upon UV irradiation, the membrane permeance was reduced to 19.3 L/m²/h/bar in the cis state. The process of photoisomerization facilitates a significant geometrical transformation in azobenzene compounds. Specifically, it transitions from the planar trans-azobenzene form, which measures ~9 Å, to the nonplanar cis-azobenzene configuration, which is about 6 Å in size. In the CON, the cyclen ring is subjected to tension, causing it to transition from a zigzag conformation to a planar conformation. This alteration results in corresponding size modifications from 3.4 Å to 4.5 Å, while the azobenzene unit experiences a reduction in size from 9 Å to 6 Å. This critical transformation allows for the direct conversion of light energy into various mechanical motions, such as bending, oscillation, and twisting. These mechanical phenomena play a crucial role in altering the structural dynamics of materials, leading to the expansion and bending of pores within a membrane. Consequently, this alteration affects the membrane’s permeation capability. The isomerization of light-responsive azobenzene units provides a larger geometric change. In CON membranes, this light-responsive characteristic allows pores to switch between open and closed states, allowing for accurate molecular sieving of dyes while maintaining high solvent permeability. Under UV light, the cis form (off state) produces smaller pore diameters, leading to greater dye rejection than the trans form (on state), which has larger pores.

Fig. 2

Schematic illustration of the trans-to-cis and cis-to-trans photoisomerization, along with the chemical structures of light-responsive membranes²³. Permission to reprint granted.

Table 2 Comparison of different photo-responsive membranes with their operational wavelengths (activation λ_A and deactivation λ_d) given in nm, membranes fabrication, pore size change during light exposure, and practical application

The trans-to-cis photoisomerization occurred rapidly within 5 min of UV exposure, while the reverse cis-to-trans isomerization required around 30 mins under visible light, enabling reversible remote control over molecular sieving capability. Light-responsive (PAN/Azo-MPEG/Azo-PHMB) thin-film membranes were developed to enhance separation efficiency and improve fouling resistance (Table 2, entry 2)²⁴. The performance of the PAN/Azo-MPEG/Azo-PHMB membrane was outstanding, with a high-water permeability of 17.9 L/m²/h/bar and an impressive flux recovery ratio (FRR > 90%) even after multiple antifouling tests. Additionally, the membrane demonstrated high selectivity, particularly favoring divalent over monovalent salts, with a selectivity ratio of αMgSO₄/NaCl = 33.4. The functional layer of the membranes could be regenerated using ultraviolet (UV) light, ensuring sustained performance even after contamination. Furthermore, the membrane showed high rejection rates for the antibiotics erythromycin (ERY) while allowing significant passage of NaCl, making it ideal for applications in antibiotic separation from salt/antibiotic mixtures. Even after the fourth cycle, the membrane maintained low flux loss and high flux recovery, which can be attributed to the functional polymers’ hydrophilic nature and spatial repulsion properties.

The development of self-cleaning ultrafiltration (UF) membranes was achieved by co-depositing photo-mobile 4,4′-azodianiline (AZO) and bio-adhesive polydopamine (PDA) (Table 2 entry 3)²⁵. These membranes show photoresponsive behavior in the presence of UV/Vis light irradiation. This process mitigated foulant aggregation and significantly enhanced the surface hydrophilicity, demonstrated by a decrease in water contact angle (WCA) from 39° to 29°. The combination of PDA and AZO in the coating layer provided robust adhesion to the membrane surface while preventing pore blockage, ensuring high water permeance. The modified membrane demonstrated outstanding self-cleaning ability, with a 160% increase in water permeance following UV/visible light exposure, even when contaminated with bovine serum albumin (BSA). This strategy presents a promising pathway for creating membranes with better fouling resistance and long-term, sustainable performance, especially in water purification applications that require frequent cleaning.

The thin-film membranes (SulfAzo3TB/PES) were designed to provide light-responsive control over water permeability and molecular separation (Table 2, entry 4)²⁶. These membranes exhibited a high-water permeability, with a permeance variation of 0.78 L/m²/h/bar in the cis state and 0.64 L/m²/h/bar in the trans-state, demonstrating their reversible tunability. The membranes showed selective molecular weight cut-off (MWCO) variation, shifting from 2000 g/mol in the trans-state to 4000 g/mol in the cis state. The light-responsive properties of the membranes enabled easy regeneration through UV/visible light exposure, maintaining high performance and permeability over multiple cycles. Furthermore, the membranes demonstrated high rejection rates for larger organic molecules like Rhodamine B, while allowing smaller molecules to pass, making them ideal for applications requiring size-selective filtration. The light-responsive switching ability of azobenzene has been used effectively for oil-water separation. In this method, azobenzene groups were attached to SiO₂-modified polypropylene (PP) membranes through a straightforward chemical grafting process (Table 2, entry 5)²⁷.

The grafted chromophore, 7-[(trifluoromethoxyphenylazo)phenoxy]pentanoic acid (CF₃AZO), adopts the thermodynamically stable trans configuration under ambient conditions, imparting a super-hydrophobic WCA of ≈160°. Exposure to 365 nm UV light drives a rapid trans → cis photo-isomerisation; the higher surface free energy of the cis form collapses the WCA to ≤5°, producing a super-hydrophilic surface. Re-illumination with visible light at 440 nm restores the trans-state and the original super-hydrophobicity, completing a fully reversible cycle that can be repeated for many hundreds of switching events without fatigue. This membrane delivers outstanding performance in separating oil from water, with 98.7% efficiency 98.7% efficiency and a high flux of 2352 L/m²/h, which increases to 2653 L/m²/h for oil permeation under visible light. It exhibits excellent physical and chemical stability, retaining its performance in harsh conditions and after repeated mechanical stress and rinsing. Its mechanical flexibility and durability make it an ideal material for sustainable and efficient industrial oily water separation.

To further improve self-cleaning and desalination properties, the thin-film composite membranes (0.10%-PPy@G-CN/HCPAM) were developed (Table 2, entry 6)²⁸. These membranes demonstrated excellent performance, achieving a high-water flux of 78.57 L/m²/h at 25 bars with nearly complete rejection of Eriochrome Black T (EBT) dye at 99.90%. Another light-responsive nanofiltration membrane was developed using a covalent organic framework composite (β-CD/AZO COF/HPAN) membrane, incorporating azobenzene (AZO) functional groups to create light-gated molecular channels (Table 2, entry 7)²⁹. The membrane performance improves under 365 nm UV irradiation, where the permeance increases from 19.56 L/m²/h/bar to 42.30 L/m²/h/bar, and the dye rejection of Rose Bengal (RB) rises from 94.40% to 99.12%. The UV-triggered trans-to-cis isomerization of AZO leads to adjustable pore sizes, enhancing the membrane’s molecular sieving capabilities. The membranes exhibit excellent antifouling properties, with a fouling recovery rate (FRR) improving by 17% under UV exposure related to the higher hydrophilic nature and reduced pore size of the membranes.

Light-responsive azobenzene molecules were also introduced into the pores of a COF membrane made from 1,3,5-triformylbenzene (Tb) and 4,4′-diaminobenzanilide (Da), as shown in Fig. 3A (Table 2, entry 8)³⁰. The resulting light-responsive COF membranes, with highly ordered and fine TbDa-Azo channels, demonstrated efficient sieving ability for monovalent and multivalent ions by acting as a gate to achieve selective cation filtration. In the trans conformation, the Azo groups contributed to a pore size of ~1.07 nm, while in the cis conformation, it is ~1.42 nm. TbDa-Azo COF membranes are durable in prolonged permeation of ions (Fig. 3B). When exposed to 365 nm and 450 nm light, the membrane becomes impermeable because the pore size of the TbDa-Azo COF membrane roughly matches the diameter of hydrated Al³⁺ ions. Deformation of the hydration shells permits Al³⁺ to enter the membrane pores (Fig. 3C). The TbDa-Azo COF membrane’s water flux can be shifted between ~3.1 and ~7.5 L/m²/h under exposure to light with different wavelengths. The fine-tuned separation of the penicillin/Al³⁺ mixture is attained upon gradual and short-time exposure to UV light, and the retention of Al³⁺ is triggered by the membrane under visible light radiation for a short duration of 3 min (Fig. 3D). Azobenzene groups were also grafted onto PMMA, which was then incorporated into a P(VDF-CTFE) polymer backbone (Table 2, entry 9)³¹. The resulting membrane exhibited reversible changes in pore size and surface hydrophilicity in response to light, with UV irradiation causing pore expansion and increased hydrophilicity, thereby significantly improving backflushing efficiency. The light-triggered process recovered over 90% of irreversible fouling, such as BSA, allowing for precise molecule release based on size. Azobenzene-grafted GO-PVDF membranes exhibited light-responsive behavior in the presence of UV/Vis light (Table 2, entry 10)³². The trans membranes exhibited an MB permeance of 132 L/m/h/bar and achieved near-complete rejection of methylene blue (MB) dye, attributed to size exclusion mechanisms. In another study, the light-responsive molecules were attached to alumina membranes using a covalent condensation reaction (Table 2, entry 11)³³. In this study, MB dye was completely rejected in the trans-state, while in the cis-state, it was not completely rejected due to the size of the molecules.

Fig. 3: Synthesis of photo-responsive azo-compound and their molecular and ionic transport studies.

A Synthesis of 1,3,5-triformylbenzene (Tb)–4,4′-diaminobenzanilide (Da) compound (TbDa) and TbDa-Azo compound. B Representation of the photo induced mass transport behavior. C Multistep manipulation of AI separation factor applying UV/Visible light. Adapted with permission from ref. 30. Permission to reprint granted.

A light-responsive spiropyran (SP) moiety was incorporated into polyamide thin-film composite nanofiltration (NF) membranes in a single-step process, facilitated by low-energy electron beam treatment (Table 2, entry 12)³⁴. Another work integrated SP onto graphene oxide-based NF membranes (Table 2, entry 13)³⁵. When exposed to UV radiation, SP converts into zwitterionic merocyanine, resulting in water permeation rates of ~6.5 L/m²/h/bar. Exposure to visible light reverted merocyanine back to SP, resulting in a water flux of ~5.25 L/m²/h/bar. These modified NF membranes outperformed unmodified membranes in terms of chlorine tolerance and normalized water flux, while maintaining ion rejection capability.

SP-based membranes are produced by grafting poly(methacrylic acid) (PMAA) brushes onto polypropylene (PP) substrates via argon plasma-induced free-radical polymerization (Table 2, entry 14)³⁶. Afterward, SP moieties are introduced into the polymer brushes through post-polymerization modification. The light-responsive membranes exhibit reversible changes in wettability and permeability in response to light stimuli. The PMAA-SP-modified membranes demonstrate efficient switching properties, with the SP groups enabling a transition between hydrophobic and hydrophilic states under UV and visible light, respectively. The pore sizes are reduced in the hydrophobic state, while in the hydrophilic state, the pores expand. These membranes maintain durability and performance over prolonged cycles of stimulus exposure. The water flux through the membranes can be adjusted between ~256 and ~405 L/m²/h/bar under alternating light conditions. SP molecular aggregates into graphene oxide (GO) structures exhibit significant potential in water purification due to their tunable permeability, enhanced separation performance, and self-cleaning capabilities (Table 2, entry 15)³⁷. Incorporating SP aggregates enriches water transport channels and increases hydrophilicity, resulting in exceptional water fluxes, such as 95.0 L/m²/h/bar under visible light conditions. These membranes demonstrate superior rejection rates for various dye molecules, driven by size exclusion and electrostatic interactions³⁸. Upon light exposure, the membrane transitions from hydrophobic to hydrophilic, reversing the fouling process. This mechanism allows the membrane to recover up to 96.2% of its initial permeability after fouling by organic pollutants like bovine serum albumin. The SP was also attached to PVDF via the NIPS method (Table 2, entry 17)³⁹. Incorporating SP aggregates enriches water transport channels and increases hydrophilicity, resulting in exceptional water fluxes, such as 104 L/m²/h/bar under visible light conditions. These membranes maintain durability and performance over prolonged cycles of stimulus exposure.

Anthracene-based membranes are produced using poly(styrene-block-anthracene-block-methyl methacrylate) triblock copolymer through a self-assembly non-solvent induced phase separation method (Table 2, entry 18)⁴⁰. The light-responsive anthracene groups undergo conformational changes upon light irradiation, which enables the membranes to exhibit efficient water flux and solute retention control. Under UV light at 365 nm, the anthracene groups form [4 + 4] cycloadducts, decreasing water flux from 310 to 250 L/m²/h/bar. This reduction occurs because the cycloadducts partially close the pores of the membrane. As expected, reversing the light exposure at 264 nm restores the original flux to 280 L/m²/h/bar. The highly ordered channels of the anthracene-based membranes act as gates, achieving selective permeability. The photo-dimerization and its reverse process happen quickly under UV light exposures, making these membranes highly responsive and durable over multiple irradiation cycles.

Light-responsive materials for membrane development and future directions

This section elucidates the comparison of different light-responsive molecules in terms of isomerization time (responsiveness) based on activation energies, structural modification, photostability, and processability behavior.

Based on activation energies, azobenzene is known for its fast photoisomerization between its cis and trans forms, occurring in just a few seconds due to its low activation energy (~50 kJ/mol). In contrast, diarylethene’s ring-opening and closing reactions take about 1 min due to a higher activation energy (~100 KJ/mol). Spiropyrans and stilbene exhibit slower responsiveness due to complex isomerization mechanisms and higher energy barriers¹³. Anthracene’s photodimerization process, requiring ~200 kJ/mol, is even slower, taking 20–60 mins and resulting in a much longer response time compared to other light-responsive molecules¹⁴.

Structural modification of light-responsive molecules is a key strategy for developing light-responsive membranes. Structural modification includes changes to the molecular backbone, the addition of functional groups, or alterations of side chains, each affecting light-responsive properties in distinct ways. For instance, azobenzene absorbs UV-A light in the 320–380 nm range and exhibits rapid photoisomerization due to low activation energy, which allows it to respond quickly¹⁵. Extending conjugated systems in azobenzene structure can shift absorption wavelengths (365 nm), isomerization kinetics and activation energy, thus controlling the water flux⁴¹. These structural adjustments influence the efficiency and wavelength of the light response and affect how the molecule interacts with the membrane matrix, impacting solubility, phase separation, and mechanical properties⁴². Spiropyrans also respond to UV-A light (350–400 nm) but have higher activation energy, resulting in slower isomerization¹⁶. In contrast, anthracene absorbs at 315–360 nm, needing even more energy for photodimerization, while stilbene, absorbing in the 280–320 nm range, shows lower responsiveness due to its complex isomerization mechanisms¹⁷. All these molecules show responsiveness under UV-A light, but the differences in activation energy led to varying reaction rates. The density of the functional groups is also a crucial strategy for enhancing the performance of light-responsive membranes. Higher concentrations of light-responsive units can lead to increased responsiveness and faster switching times, resulting in more pronounced changes in permeability upon light exposure. The density of functional groups can also influence the membrane microstructure and phase behavior, allowing for tailored permeability profiles and dynamic gating⁴³.

Molecular size variation is another important strategy in designing photo-responsive membranes, significantly influencing their responsiveness, properties, and overall performance³⁷. The size of light-responsive molecules is crucial for determining membrane effectiveness, as it affects their interaction with light and subsequent changes in membrane characteristics⁴⁴. Smaller molecules allow for denser packing within the membrane matrix, enhancing light responsiveness and enabling rapid changes in permeability upon irradiation. Increasing the density of light-responsive molecules also influences membrane performance. Larger photo-responsive units provide structural stability and strength, which helps maintain functionality under varying operational conditions. However, these larger units may hinder the mobility of light-responsive sites, potentially resulting in slower response times. This relationship underscores the need for careful consideration in membrane design to balance responsiveness and structural integrity.

The multi-layer approach is an effective strategy for developing advanced light-responsive membranes by integrating various materials and functionalities. This approach involves stacking different photo-responsive materials, each selected for its unique properties and responsiveness to specific wavelengths of light. For example, a top layer of azobenzene can facilitate rapid switching and permeability changes, while a bottom layer of spiropyran enhances wettability and fouling resistance. This method improves mechanical stability and allows for the integration of light responsiveness with other external stimuli, like pH or temperature.

Photostability is another important parameter assessing the durability and effectiveness of light-responsive molecules. Diarylethene exhibits exceptional photostability in its closed-ring form, where it shows strong resistance to photo fatigue over many cycles (35 cycles) of light exposure¹⁸. On the other hand, azobenzene shows lower photostability due to UV light-induced photodegradation and thermal back-isomerization¹⁹. Anthracene, however, demonstrates high photostability through a reversible photodimerization process, where two anthracene molecules form a covalent bond. This dimerized form is stable and resistant to photodegradation, making anthracene suitable for applications that require long-term stability under UV light²⁰. Stilbene and spiropyran exhibit lower photostability due to environmental sensitivity (pH and solvent polarity) and photodegradation, which reduces their durability under UV light exposure²¹.

The processability of light-responsive molecules varies significantly depending on their chemical structure and physical properties. Azobenzene materials are more processable and can be easily incorporated into various polymer matrices and other materials, creating materials with tunable properties^22,23. For instance, azobenzene has the ability to form host-guest complexes with cyclodextrin; the stability of these complexes is dependent on the isomeric state of azobenzene²⁵. This light-triggered isomerization allows for reversible switching between complexed and uncomplexed states, which has been used to build a variety of light-responsive materials²⁵.

In contrast, spiropyran, while also processable, requires more careful handling due to its sensitivity to pH and solvent polarity, which can affect its functionality during processing²⁶. Diarylethene is also challenging to process due to its rigid closed-ring structure, needing specialized techniques for dispersion in polymers, though it can enhance electro-switch characteristics when incorporated into conjugated polymers²⁷. Lastly, anthracene’s slower photodimerization can lead to inefficiencies in manufacturing when rapid response times are desired²⁸.

The development of light-responsive membranes critically depends on the rational design and structural modification of the photoactive molecules embedded within them. By tailoring the molecular architecture—such as the backbone rigidity, functional group density, and side chain composition—key properties including switching kinetics, activation energy barriers, and overall responsiveness can be precisely modulated. These parameters directly influence the membrane’s ability to undergo rapid and reversible transformations in permeability or surface wettability upon light exposure. Importantly, not only the molecular structure but also the spatial organization of light-responsive units within the membrane matrix plays a vital role. A periodic or ordered arrangement of these functional groups is expected to promote cooperative transitions and more efficient signal propagation across the material, whereas a random distribution may lead to diminished or heterogeneous responsiveness. This highlights the importance of supramolecular engineering strategies in membrane fabrication. In parallel, a major research focus is on creating inherently photo-responsive polymer systems that can be directly processed into membranes, eliminating the need for post-synthetic surface functionalization. Such materials would enable higher functional group density and uniformity, while simplifying the fabrication workflow. However, identifying polymers that combine sufficient photo-responsiveness with mechanical integrity and processability remains a significant challenge.

To translate these materials into real-world applications, light-responsive behavior must remain effective at larger membrane scales and under practical operating conditions. The switching performance must be stable over multiple light on/off cycles, and compatible with scalable manufacturing techniques such as roll-to-roll or layer-by-layer assembly. Potential applications include tunable membranes for selective molecular separation, on-demand drug delivery systems, light-triggered valves in microfluidic devices, and self-cleaning surfaces in water treatment modules. Future work should therefore focus on aligning molecular design with application-specific performance metrics and scalable fabrication methods.

Photo-responsive membranes for water treatment report high solar efficiencies alongside variable water flux and cost-performance metrics. Composite membranes using Ti₃C₂Tx MXene achieve water evaporation efficiencies of about 90%, while Janus membranes deliver ~92% solar energy utilization. Au nanoparticle-based composites reach up to 94.6% solar thermal conversion, and polydopamine-coated systems and carbon nanotube membranes exhibit light absorption values near 97% and 94%, respectively.

Cost information appears less consistently across studies. When measured, performance ratios include 62.35 g·h⁻¹ per dollar and $0.15 per kg·m⁻²·h⁻¹, while minimal use of expensive materials in AuNP systems is noted. In sum, these studies indicate that high solar efficiency and acceptable cost performance can be obtained with photo-responsive membranes, though differences in reporting preclude a fully standardized cost comparison. On a cost comparison basis, it can be said that high solar efficiency is achievable with a range of photo-responsive membrane materials, including those using minimal amounts of expensive components. However, only a few reports have provided quantitative cost data or cost-performance ratios. The available evidence suggests technical feasibility and potential for cost-effectiveness, but the lack of consistent economic and operational data precludes firm comparative conclusions.

Although, light-responsive materials offer great potential for the controlled release of substances from light-responsive membrane matrices. However, challenges remain, particularly with the availability of compatible light-responsive materials and two-dimensional materials for composite membrane fabrication. Additionally, the use of light may increase operational costs. Despite these hurdles, there are significant opportunities to improve the large-scale manufacturing and practical applications of such smart membranes with antifouling and self-cleaning properties. Key areas for improvement include enhancing photo-responsive efficiency and stability. Optimizing the membrane’s pore structure can strengthen interactions for better rejection and permeability. Further research on light-responsive membranes in photocatalytic membrane reactors can assess their stability and self-cleaning capabilities. Additionally, exploring pollutant removal mechanisms and how the positioning of the membrane module influences the effectiveness of light could enhance performance.

Electro-responsive materials (ERMs)

When subjected to external electrical impulses, electro-responsive materials can reversibly alter their microstructure, resulting in changes to their physical and chemical properties and enabling controlled, tunable performance in various applications. This section will explore the chemistry and adjustable properties of electro-responsive materials, including electrical conductivity, mechanical properties (such as swelling/stiffness), and conformational changes, which have been used to date for responsive membrane development. It will underscore both the challenges and improvement opportunities.

Electro-responsive polymers (ERPs)

Electro-responsive polymers (ERPs) possess the capability to transduce electrical energy into mechanical energy, such as swelling, shrinking, or bending under an applied voltage. They are a class of inherently conductive polymers with π-electron delocalization in their backbones, including polyanilines (PANIs), polythiophenes (PTs), and polypyrroles (PPys)^45,46. A detailed description of their individual chemistries and electroresponsive behavior mechanisms is provided in the supporting information.

Although most ERPs share similar redox mechanisms, the total volume change of an ERP during redox switching depends on several interrelated effects, including change in charge density, electronic response of an electric double layer at the polymer-electrolyte interface, and osmotic expansion/redox interaction between electrolyte and ERP⁴⁷. Changes in the charged density along the conductive polymer backbone result in the deformation of the polymer network, as the C-C bond length is altered, as shown in Figs. S3 and S4. This volume change occurs due to fluctuations in charge density during processes of oxidation or reduction. The electric double layer at the polymer-electrolyte interface reacts to alterations in the electronic state, also leading to conformational changes.

Osmotic expansion in electro-responsive polymers (ERPs) is caused by redox interactions between the conducting polymer and electrolyte molecules, which promote solvent flow and change the volume of the ERP matrix⁴⁷. ERP actuation in electrochemical redox processes is enabled by the movement of mobile ions: in anion-driven systems, the polymer backbone loses electrons while anions are incorporated; in cation-driven systems, the backbone gains electrons and cations are taken up, as illustrated in Figs. S3A–C and S4^47,48,49.

Carbon-based materials and single molecules

Electro-responsive membranes made with carbon-based materials exhibit excellent conductivity due to their sp²-hybridized carbon atoms and delocalized π–π electron systems. These materials can function as either cathodes or anodes. For applications such as electrostatic repulsion, electrophoresis, and direct redox reactions, voltages below the oxygen evolution threshold of carbon electrodes are generally used⁵⁰.

Carbon anodes have a potential for the oxygen evolution reaction generally below 0.4 V. They interact with electrogenerated hydroxyl radicals to partially oxidize organic compounds. They interact with electrogenerated hydroxyl radicals to partially oxidize organic compounds. Carbon nanotubes (CNTs) are commonly employed as conductive materials in electro-responsive membrane manufacturing due to their ability to create dense, interwoven structures with porosity and support high current densities of over 10⁹A/cm² for metallic single-walled CNTs⁵¹. This structural configuration facilitates inter-tube bonding through van der Waals interactions, ensuring operational stability. Furthermore, graphene-based carbon materials also manifest advantageous electrical properties. The synthesis of a conductive layer through deposition results in a specific surface area and tunable pore size distribution. Introducing oxygen-containing groups allows CNT and graphene-based materials to interact with other metal catalysts, such as Pd and Fe. Moreover, mesh-type electrodes like carbon fiber cloth and carbon paper offer effective conductive platforms for membrane modification⁵².

MXenes, with their distinctive layered structure composed of alternating transition metal and carbon layers, are particularly attractive for electrocatalysis and energy storage fields. Their high electrical conductivity reaching up to 10,000 S/cm in the case of Ti₃C₂Tₓ enables rapid electron transport across the electrode material⁵³. The surface functional groups and high specific surface area of MXenes provide numerous sites for anchoring active catalysts. Additionally, the surface of MXenes behaves like a transition metal oxide, making it redox-active, encouraging their use in electro-responsive membranes. However, during electrode fabrication, MXene nanosheets tend to cluster and self-assemble due to strong hydrogen bond interactions and van der Waals interactions, which can compromise their electrochemical performance. Single molecules that are electro-responsive can have their molecular energy levels adjusted when the gate electrode potential is maintained beyond the boundaries of the redox-active range. Upon entering the redox region, the molecules experience reversible electron transfer processes, which modify their energy states and the extent of conjugation⁵⁴. Viologen and its derivatives are recognized for their conjugated structures, which enable electrical conductivity. The nitrogen atoms in their pyridine rings provide lone pairs that enhance electron transport. These molecules can cycle through three different oxidation states by reversibly transferring two electrons, making them valuable in technologies such as supercapacitors and redox flow batteries, where their response can be finely tuned by an applied electric field.

Viologens exist in various states, including dicationic, monocation radical, or redox couple, to meet diverse energy system requirements. The viologen dication interacts electrostatically with polysulfides in Li–S batteries, while its radical form helps inhibit lithium dendrite growth in Li-ion batteries⁵⁵. Their redox behavior is advantageous in redox flow batteries, Li-air batteries, and supercapacitors, contributing to enhanced performance. This switchable redox state feature endows viologens and their derivatives with excellent cyclic reversibility, supporting their role in membrane development⁵⁶.

These carbon-based frameworks and molecules also offer a unique platform for constructing electro-responsive membranes with tunable transport properties. When an external electric field is applied, charge redistribution at the conductive interface can dynamically alter the membrane’s surface charge density, wettability, and ion transport resistance. Furthermore, covalent or non-covalent integration of redox-active single molecules—such as viologens, ferrocene derivatives, or metal-organic complexes—onto CNT or graphene scaffolds allows for precise modulation of membrane behavior via reversible electron transfer processes. This integration enables programmable gating, selective ion rejection, or antifouling switching, making carbon-based systems ideal for next-generation membranes in separations, environmental remediation, and biointerfaces.

Electro-responsive enabled membranes

The ERMs can be integrated into the membrane matrix, chemically grafted to its structure, or coated onto the surfaces and pores of the membranes as presented in Fig. 4^57,58. This section will discuss how ERMs are incorporated into membranes to control permeability, rejection, wettability, catalytic degradation, and pore size under applied potential. Table 3 provides an overview of recent advances in electro-responsive membranes, however several challenges still need to be overcome.

Fig. 4: Electro-responsive membranes (ERMs) for water filtration and purification.

A When a voltage is applied, the ERM interacts with contaminants through electrooxidation, electrostatic adsorption, and electrostatic repulsion. Electrooxidation is primarily responsible for water decontamination⁵⁰. B The redox-responsive membrane transitions from a hydrophobic to a hydrophilic state upon the application of a voltage. This change facilitates the controlled transport of molecules and ions across the membrane¹⁶⁶. Permission to reprint granted.

Table 3 Electro-responsive membranes for different applications

Polymer-based electro-responsive membranes

The electrical regulation of oxidation and reduction properties of HCl-doped PANI membranes containing different molecular weights of PANI in relation to the applied potential was studied⁵⁹. In aryl amine polymers, when the pH is very low, anions enter the film during the oxidation and exit during the reduction. This movement may also carry solvent molecules, resulting in swelling and deswelling of the polymer⁶⁰.

Applying a high potential can modify the conjugated framework by inducing oxidation and triggering the dopant migration, changing their binding sites or spatial arrangement within the polymer, and thereby modifying the membrane’s free volume. This morphological shift causes polymer chains to swell, in turn reducing the membrane’s free volume and pore size. As a result, this causes additional resistance to water transport, leading to a change in permeance at different applied voltages (Table 3, entry 1). The PANI membrane’s regulation mechanism is influenced by Donnan repulsion, which depends on surface charge, as well as by diffusion processes determined by its chemical characteristics. Acid or alkali doping can also control the porosity of PANI membranes. However, conventional small-acid doped PANI membranes offer limited chemical resistance and mechanical strength, limiting the stability, durability, and nanofiltration applications of membranes⁶¹. It was shown that exposure to acids may strongly weaken intermolecular interactions and eventually embrittle the membrane material^61,62. To address these issues, the researchers opted for poly(2-acrylamido-2-methyl-1-propanesulfonic acid) (PAMPSA) as the acid dopant. In contrast to HCl, PAMPSA is a flexible polymeric acid with a bulky benzene sulfonyl group. SO₃H groups are covalently linked to benzene rings, which are responsible for stable hydrophilic modification and doping of the system. The strong covalent bond between the acid group and the benzene ring reduces acid leaching, while the dopant functions as a plasticizer, increasing the membrane’s mechanical strength.

The electrical conductivity of the PAMPSA-doped membrane showed no change when tested both before and after PEG filtration. In contrast, the conductivity of the HCl-doped membrane decreased significantly by four orders of magnitude following PEG filtration. This substantial difference in conductivity suggests that there is minimal leaching of the dopant in the PAMPSA-doped membranes. Electrically induced swollen polymer chains reduced the rejection of PEG to 32% for the PAMPSA-doped PANI membrane and to 95% for the HCl-doped membranes. The lesser rejection for PAMPSA-doped membranes indicated that the PANI-PAMPSA membrane has a more open structure than the PANI-HCl membrane. Larger dopants, such as PAMPSA, widen the intermolecular spacing among PANI polymer chains 64⁶³. As a result, this facilitates the development of a looser membrane structure with larger pore sizes (Table 3, entry 2).

Nano-based electro-responsive membranes

One effective technique to enhance the rejection efficiency of the NF membrane involves reducing pore size or strengthening electrostatic interactions. However, this strategy has a disadvantage: smaller pore sizes increase rejection while decreasing permeability, which might reduce overall separation efficiency. In contrast, increasing electrostatic interactions might result in high rejection rates while maintaining permeability⁵⁶.

Incorporating polystyrene (PSS) into the PANI@CNT membrane can significantly increase its surface charge density, increasing it from 11.9 to 73.0 mC/m²⁵⁷. The increased surface charge density under electrical assistance leads to improved ion rejection rates for both Na₂SO₄ (from 81.6% to 93.0%) and NaCl (from 53.9% to 82.4%), without sacrificing the water permeability of the membrane (Table 3, entry 3). The electrical assistance enhances the Donnan potential difference between the membrane and the bulk solution, which increases the ion transfer resistance and thus improves the ion rejection performance. Another group of researchers evaluated the membrane permeance and ion rejection performance of PANI-DBSA-CNT and PANI-PSSA-CNT membranes for NaCl and Na₂SO₄ under different applied electrical potentials⁵⁸. PANI-PSSA-CNTs showed the highest surface charge density due to the PSSA dopant, which has a higher degree of deprotonation compared to DBSA (Table 3, entry 4).

A PANI-entangled oxidized CNTs (PANI@O-CNTs) membrane was also developed⁵⁷. The PANI@O-CNTs membrane exhibited superior conductivity and stability due to in-situ polymerization of the PANI network on the CNTs framework. When an external voltage was supplied to this membrane, rejection rates for both the negatively charged methyl orange (MO) and positively charged methylene blue (MB) dyes increased dramatically (Table 3, entry 5). After 50 minutes of operation at a voltage of -2.0 V, the membrane’s rejection of MO increased from 35.2% to 77.4% and of MB from 41.5% to 79.8%. Combining electrostatic interactions with H₂O₂ oxidation led to higher rejection rates of 89.7% for MO and 93.4% for MB. Additionally, combining electrostatic interactions with H₂O₂ oxidation led to higher rejection rates of 89.7% for MO and 93.4% for MB.

The effect of oxygen plasma treatment was also investigated on PANI membranes made with carbon nanofibers (CNFs). Microfiltration PANI@O-CNF membranes were created by electrospinning PANI, followed by carbonization and oxygen plasma treatment⁶⁴. The synthesis of PANI@O-CNF membranes improved the wettability and electrocatalytic reactivity for MB and acetaminophen (ACP) degradation (29.6 × 10³ min^-1 electrocatalytic constant for MB and 15.6 × 10³ min^-1 electrocatalytic constant for ACP) (Table 3, entry 6). Under combined microfiltration and electrocatalytic conditions, the removal efficiency of MB and acetaminophen reached 99% and 91%, respectively. Increased rejection after plasma treatment was due to strong electrostatic interactions between the negatively charged sites formed on the O-CNF membranes and the cationic MB dye molecules, while hydrophilic interactions with oxygen-containing functional groups on the membrane surface also contributed to increased ACP elimination.

The oxygen evolution potential (OEP) of pristine CNF membranes increased from 1.29 V to 1.62 V after 5–7 minutes of plasma treatment, which is related to improved electrocatalytic performance of the treated membrane. Recent work has developed a PPy@OCNT membrane by polymerizing PPy onto OCNT-modified PVDF substrate⁶⁵. The developed membrane experienced an 82% drop in overall fouling resistance and a 95% decrease in irreversible fouling under applied potential, confirmed by electrically regulated hydraulic backwash. This improved the membrane permeance recovery rate from 39.6% without voltage to 99.1% at +1.19 V (Table 3, entry 7). A new approach to create conductive nanofiltration membranes achieved a 90% improvement in permeability, reaching 20.4 L/m²/h/bar, by combining interfacial polymerization with in-situ self-polymerization of EDOT (Table 3, Entry 8)⁶⁶. Despite PEDTOT being a well-known conductive polymer, its poor solubility makes it challenging to use in membrane fabrication.

The semi-flexible π-conjugated backbone of PTI promotes polymer chain aggregation in solution, resulting in limited solubility, reduced mechanical flexibility, and challenging processing properties. In contrast, the EDOT monomer has good solubility properties. By combining interfacial polymerization with in situ self-polymerization of EDOT, the solubility issues of PEDOT can be solved, resulting in stable, highly conductive nanofiltration membranes. Using these PEDOT-doped membranes with an electro-assisted cleaning approach achieved quick flux recovery to 98.3% in just 5 minutes, significantly surpassing the standard 30-minute pure water cleaning. The anticipated electricity cost is only $0.055 per day.

A facile modulation of water transport through laminar Ti₃C₂Tx MXene membrane using an electric field was reported⁶⁷ Fig. 5A (Table 3, entry 9). The permeability of water through a laminar MXene membrane has been significantly improved up to 70 times under a negative voltage of −5 V, while maintaining a 91% and 94% rejection for Cango Red and aniline, respectively. To prevent restacking of MXene nanosheets and improve water transport pathways. MXene@CNT nanofiltration membrane displayed 96% and 87% rejection for orange G and MO dyes under applied potential of 3 V (Table 3, entry 10)⁶⁸. The enhanced water permeation and high dye rejection observed in MXene-based electro-responsive membranes result from field-induced expansion of interlayer spacing via electrostatic repulsion, suppression of nanosheet restacking by CNT spacers, and dynamic gating effects. Applied voltage modulates nanochannel architecture and surface potential, enabling increased water flux and selective ion/dye exclusion. These synergistic mechanisms establish MXene-based systems as capable runners for tunable, high-performance separation membranes. In another study, CNT-based ultrafiltration membranes were introduced using solid-state dry spinning of drawable CNTs onto carbon nanofiber (CNF) support membranes by varying the number of CNT layers (5, 10, 20, 30, 60) and the orientation angle of the CNT layers (0°, 45°, 90°)⁶⁹. The achieved reaction kinetic constant for MB and ACP was 1.1 to 3.9 times higher than previously reported values, with a degradation efficiency of 99% for MB and ACP (Table 3, entry 11). The electrostatic interaction between the polar C–F bonds in PVDF chains and the mobile π-electron cloud of graphene promotes their self-assembly. This results in a strong β-phase, giving the composite film unique piezoelectric capabilities, enabling precise molecule separation. When a voltage is applied, the enlarged free volume fills the gaps between graphene and PVDF, forming nanochannels through which only particular molecules can travel. This mechanism reduces total permeability while boosting CO₂ selectivity (Table 3, Entry 13)^70,71.

Fig. 5: Voltage-induced changes across different membranes.

A Digital top-view image of a Ti₃C₂Tx membrane on supporting PVDF filter paper & Molecular model of Ti₃C₂Tx showing two layers of water with ions between the layers. Adapted with permission from ref. ⁷², Copyright (2018) American Chemical Society. B Reversible change of pore size between oxidation and reduction states of DBS-doped PPy membrane. Adapted with permission from ref. ⁷⁵. Copyright (2011) American Chemical Society. C Water permeation for PPy-DBS membrane in the presence of applied voltage⁷⁴ and D the electro-responsive gated mechanism based on a reversible wettability switch of the PFOS-doped. PPymembrane⁷⁷. Permission to reprint granted.

Electro-responsive membranes with gated functionalities

The utilization of electrically responsive membranes for voltage-dependent exclusion mechanisms for ions and molecules has been proposed as an effective method to improve membrane selectivity, increase rejection efficiency, and enhance operational stability. Over the past few decades, various materials, including polymers, CNTs, MXenes, and viologen derivatives, have exhibited significant gating effects on molecules when subjected to an applied potential⁷². An electrically switchable membrane was developed by electrochemical polymerization of AOT-doped-PPy onto CNT-deposited PVDF support⁷³.

The designed gated membrane can effectively eliminate organic micropollutants and metal ion contaminants from water by combining hydrophobic and electrostatic interactions. By providing a reversible electrical potential, trapped impurities can be released, allowing the membrane to regenerate and self-clean without the need for further chemicals (Table 4, Entry 1). The voltage-responsive gating behavior is caused by the reorientation of AOT molecules, which is influenced by the redox state of PPy under applied voltage. This transition impacts the solvent accessibility of the hydrophobic alkyl chains and negatively charged sulfonate groups of AOT, modulating the membrane’s surface charge and wettability, resulting in quick and effective single-pass separation of micropollutants and metals. This allows the hydrophobic alkyl chain and the negatively charged sulfonate group of AOT to switch positions, altering the membrane’s surface wettability and charge. As a result, efficient and rapid removal of micropollutants and metal ion contaminants from water is achieved through a single-pass filtration process. The membrane displayed switchable wettability change from superhydrophobic (145°) to superhydrophilic (13°) by oxidation (1.6 V) and reduction (−0.8 V), respectively. The AOT-PPY@CNT membrane effectively removed various organic contaminants with much lower energy consumption than a commercial nanofiltration membrane. Electrically-responsive PPY-DBS membrane with in-situ regulation of pore size can be tuned to alleviate fouling and enable selective separation⁷⁴. Applying a negative voltage to PPy-DBS causes the retention of bulky DBS ions within the polymer structure, resulting in an accumulation of negative charges across the membrane.

Table 4 Electro-responsive membranes with gated functionalities

External cations enter the PPy to compensate, increasing the distance between chains, leading to volume expansion and smaller pores. A positive potential causes cations to be expelled, resulting in volume shrinkage. Application of an oxidation potential of 0.7 V reduces the membrane’s pore size, whereas a reduction potential of −0.7 V enlarges the pore size (Table 4, entry 2) (Fig. 5B). The membrane’s specific flux in response to applied voltages was 21.9% higher than that without voltage. PPY-BDS membrane also displayed pulsatile drug release, highlighting the need for controlled and long-term drug release of protein therapeutics and the advantages of pulsatile drug delivery for specific medical conditions (Table 4, entry 3)⁷⁵. Another research explored how applying various redox potentials (−0.9/0.6 V, −0.9/0.1 V, −0.6/0.1 V) to PPy-DBS surfaces in water enables controlled detention and release of DCM droplets, examining how these voltages influence droplet retention and release behavior⁷⁶.

Results revealed that droplet release time depended on both the applied redox voltage and the thickness of the PPy-DBS layer. The quickest release occurred at −0.9/0.1 V, while increasing the coating thickness from 0.6 μm to 5.1 μm led to longer release times (Table 4, entry 4; Fig. 5C), which is due to slower desorption of DBS⁻ ions from the thicker PPy-DBS surface. The paper has not explored the capture and release of organic droplets in a systematic way, and there is room for further research on controlling the release process and understanding the effects of experimental parameters. The release time depends on the thickness of the PPy-DBS coating, which could be seen as a limitation in controlling the release time.

The longevity of the PPy-DBS surfaces is limited, as the surfaces eventually lose their ability to release the captured droplets after repeated redox cycles. Incorporation of PFOS⁻ ions into PPy micro/nanoporous film onto an anodic aluminum oxide (AAO) nanoporous membrane was demonstrated to create an electrically actuated nanochannel system for controlled, pulsatile release of drugs such as penicillin G sodium and Rhodamine B⁷⁷ (Table 4, entry 5) (Fig. 5D). The maximum pore size that can be effectively gated may be limited, as the paper notes the gating ratio decreased with further increases in pore size beyond 59 nm. The drug release performance may be affected by the charge and properties of the drug molecules, as the paper found a slower release rate for the cationic Rhodamine B compared to the anionic penicillin G sodium.

Viologen derivatives, such as cucurbit[7]uril (CB[7]), were also used for molecular nanofiltration. The viologen enhances membrane conductivity and serves as an electric switch valve, while CB[7] stops viologen dimerization and enhances its oxidation-reduction stability⁵⁶. The electric field increases repulsion between the membrane and dye, improving separation efficiency, especially for similar-sized molecules. Overall, this electrically-gated NF membrane demonstrates higher responsiveness and cycle stability compared to other advanced membranes (Table 4, entry 6).

Gated membranes based on carbon materials have recently gained significant interest for their exceptional ion-sieving capabilities. Notably, researchers have identified that precise ion sieving can be attained by deliberately manipulating the interlayer spacing of carbon-based materials. For example, CNT-based hollow-fiber membranes were used to construct a simple biomimetic membrane system that exhibits voltage-gated transport of nanoparticles across their pore channels⁷⁸. Voltage-gated transport was attributed to the noncovalent interactions between the nanoparticles and the pore channels, which are controlled by the polarity of the CNTs. While working as a cathode, when voltage was applied to CNT membranes at various levels (0.4–0.6 V), membranes behaved positively or negatively, which in turn caused rejection or penetration of GNPs of different sizes (10 and 40 nm) (Table 4, entry 7). Switchable rejection rates of different ions and molecules through Ti₃C₂Tx MXene membranes were studied by applying positive (0.4 V) and negative (–0.6 V) potentials across the membrane⁷². MXene membranes as thin as 100 nm showed rejection rates above 97% for MB dye molecules, and the rejection can be further tuned by applying negative voltages (Table 4, entry 8). The voltage-gated rejection is attributed to the control of the MXene interlayer spacing, which expands under positive voltages and contracts under negative voltages, thereby affecting ion and molecule transport. Contraction and expansion with voltage are related to the movement of ions into and out of the MXene layers, which alter the electrostatic and steric interactions between the layers. Conductivity and voltage-gated ion transport behavior of MXene membranes were further improved by incorporating GO graphene oxide within MXene⁷⁹. Applying a positive potential (+0.6 V) to MXene-GO enhances the electrostatic repulsion between the charged MXene-GO sheets and cations (Li⁺, Mg²⁺, or Al³⁺), promoting ion permeation. Applying a negative potential (-0.6 V) to MXene-GO boosts the electrostatic attraction, decreasing ion permeation (Table 4, entry 9).

Electro-responsive materials for membrane development and future directions

This section will compare different ERMs through various mechanisms, including electrostatic rejection, piezoelectric vibrations, and electrochemical reactions, which can be used to control porosity, wettability, surface charge and roughness, and catalytic efficiency for membrane water treatment applications.

Comparing the wetting characteristics of ERPs is essential for selecting the best one for membrane development. For instance, Au-coated perfluorooctanesulfonate (PFOS)-doped and de-doped PPy film displayed a superhydrophobic WCA of 152° and a superhydrophilic WCA of ~0°, respectively⁸⁰. Tetrabutylammonium hexafluorophosphate (TBAFP)-doped PTI derivative, poly(G0 − 3 T COOR), transitioned from WCA of 154° upon doping to 15° after de-doping⁸¹. For PANI, the ITO-coated surface of the PFOS-doped PANI films (at 1.05 V) showed a WCA of 153° and ⁓0° after de-doping at −0.1 V⁸². LiClO₄-doped-poly(3,4-ethylenedioxythiophene) (PEDOT) displayed reversible switching between superhydrophobic (162°) to superhydrophilic (0°) by doping and de-doping⁸³.

Comparing the wettability of carbon-based materials, graphene can switch between hydrophobic and hydrophilic states under positive and negative electric fields, highlighting the significant influence of energy on material properties⁸⁴. In contrast to carbon nanotubes (CNTs), which show minimal changes in WCA even at higher voltages, pure graphene exhibits a significant transformation, with its WCA decreasing from 117° to 86° as the voltage increases from 0 V to 40 V⁸⁵. For applications that require enhanced hydrophilicity, graphene oxide (GO) proves to be an excellent choice. Notably, GO membranes can enable precise, reversible, and electrically tunable control of water permeation by forming conductive carbon filaments that generate local electric fields. These fields induce water ionization into H₃O⁺ and OH⁻, thereby modulating flux through GO capillaries without affecting membrane integrity. This introduces a new functional paradigm for electro-responsive membranes, allowing controllable water gating without relying on external mechanical changes or complex chemical modifications⁸⁶. Additionally, MXenes, known for their exceptional 2D structure and high electrical conductivity, present exciting possibilities for electrochemical energy storage and catalysis⁸⁷. However, challenges such as self-stacking and susceptibility to oxidation highlight the need for innovative synthesis methods. Developing environmentally friendly techniques for producing high-quality MXenes is essential to fully realize their potential in membrane applications and contribute to a sustainable future.

Altering anion dopants significantly impacts electrical conductivity and surface roughness. Naphthalene-sulfonic acid dopants improve film smoothness and polymer conductivity by reorganizing polymer chains. Benzenesulfonic compounds enhance charge delocalization and electronic conductivity through strong π-π interactions with the polymer chains, facilitating charge transfer⁸⁸. The interaction improves charge delocalization in the polymer backbone, enhancing electronic conductivity through better charge transfer between chains. The presence of multiple -SO_3- groups increases conductivity, while -OH groups improve interfacial adhesion through a chelating effect. Additionally, sulfonate groups enhance intermolecular interactions and increase the dopant’s solubility via hydrophilic interactions.

The stability of ERPs plays a crucial role in determining the electrocatalytic efficiency of responsive membranes. Conductive polymers high in nitrogen, such as PANI and PPy, have been demonstrated to significantly improve membrane electrocatalytic performance. Their aromatic frameworks and strategically positioned nitrogen atoms improve polymer stability and oxidation resistance. Furthermore, the conjugated polymer backbones promote efficient electron transport, allowing for cooperative interactions with iron-based catalytic species⁸⁹.

Increased electrocatalytic performance is also related to the higher electronegativity of N atoms in the conjugated chains. This enhances the electron transfer capacity and increases the charge density of the positively charged carbon species, resulting in cleavage of the O–O bond. Conductivity is crucial for the electrocatalytic efficiency of ERPs, and based on the conductivity, PPy exhibits a higher conductivity of 105 S/cm, whereas PANI has considerably lower conductivity, which in turn lowers its catalytic efficiency⁹⁰. However, the conductivity of polymers can be enhanced through various methods, such as modifying their doping ions and levels, treating them with solvents, and adding conductive additives.

In conclusion, the application of electric fields enables the controlled release of substances from responsive membranes. However, the current challenge lies in the limited availability of easily integrable ERPs and 2D materials that are readily compatible with standard composite membrane fabrication methods. For example, it is essential to improve the electrical efficiency and cathodic electrochemical stability of the membrane while simultaneously reducing the overall costs of the device. This involves controlling the membrane pore structure to achieve strong electrostatic interactions that support high ion rejection. At higher ion concentrations, the electrical enhancement of ion rejection can weaken, necessitating the application of higher external voltages to address this issue. Enhanced precision control and tuning of the membrane’s structure to balance ion rejection and water permeance are also required. This would improve membranes’ controllability, reproducibility, and scalability based on 2D materials. Another major issue is to investigate the use of oxygen plasma-treated cellulose nanofiber (CNF) membranes in electrocatalytic membrane reactors (ECMRs), focusing on their stability and self-cleaning capabilities. Additionally, it is crucial to research how pollutants are removed and how these removal processes respond to changes in electric field intensity. We should also explore how the positioning of the membrane module within the system affects the strength of the electric field.

In summary, the application of electric fields enables the controlled release of substances from electrically responsive membrane matrices. Despite this potential, challenges remain, particularly the limited availability of ERPs and two-dimensional materials that are readily compatible with standard composite membrane fabrication methods. Moreover, the use of external electric fields can increase energy demands and operational expenses. Nevertheless, there is considerable scope for advancing the large-scale manufacturing and real-world application of electrically responsive, smart membranes with antifouling and self-cleaning capabilities.

Cost comparison analysis shows that alternative electro-responsive polymer membranes provide substantial cost advantages over traditional Nafion membranes, with alternatives costing as little as $9–47 per square meter compared to Nafion’s $700–2229 per square meter. Electro-responsive polymer membranes used in fuel cells, redox flow batteries, water treatment, and related fields exhibit significant cost differences compared to conventional perfluorinated membranes. In several studies, Nafion is reported to cost between USD 700 and 2229 per square meter. In contrast, alternative membranes—such as biochar-doped sulfonated polyether sulfone, Nafion-cellulose composites, and 3D-printed sulfonated polyether ether ketone—are described as 5–30 times less expensive, with some alternatives costing as little as USD 9–47 per square meter.

Other reports note that operational cost elements such as energy consumption and membrane replacement can represent significant cost drivers. For example, one study indicates an energy use of 0.324 kWh per kilogram in electrodeionization, while a separate investigation estimates an anion exchange membrane for carbon dioxide electrolysis at 796 euros per ton of CO produced. Polymer alternatives in redox flow batteries also achieve 99.9% capacity retention and, in one case, a 57.7% material cost reduction relative to Nafion. These findings collectively support the conclusion that, in select application areas, alternative electro-responsive membranes deliver substantial capital and performance-related cost advantages compared with traditional options.

pH-responsive materials (PRMs)

pH-responsive materials have attracted substantial attention owing to their exceptional properties to exhibit sharp and reversible changes in response to variations in external pH levels. PRMs, especially pH-responsive polymers (PRPs), demonstrate responsiveness to variations in environmental pH through modifications in their structural and property characteristics, including conformational, structural, and solubility changes. PRPs are characterized by the presence of acidic or basic functional groups that can either accept or donate protons upon external pH variations^17,91,92,93. This capability renders them highly suitable for membrane functionalization, exhibiting controlled separation performance. This section will examine the chemistry and pH-dependent characteristics of PRPs in the context of their application in membrane development, alongside future opportunities for enhancement.

pH-responsive polymers

PRPs generally utilized in membrane development are typically classified into two major types: polyacidic (polyanionic) and polybasic (polycatonic) PRPs. Polyanions feature ionizable acidic side groups along their polymer chains. In contrast, polycations include basic groups like amine (NH₂), which can be integrated into the backbone or attached as side groups. The extent of ionization depends on the pH of the environment. In contrast, polycations include basic functionalities, such as amine (NH₂) groups, which can be incorporated into the main chain or attached as side groups. The degree of ionization of these functional groups varies with the environmental pH, leading to structural changes in the polyelectrolyte⁹⁴. A polyacid tends to adopt a more extended conformation when the surrounding pH is greater than its pKa. Conversely, the polymer chains collapse when the pH is lower than the pKa. For a polybase, the situation is different: at pH values greater than its pKb, the polymer chains also collapse, while they expand when the pH is lower than the pKb. Figure S5 presents a schematic of how the ionization level of the ionic chain groups determines the states of polyacids and polybases. However, a detailed description of their individual chemistry and pH-responsive behavior mechanisms is given in the supporting information. These materials exhibit a well-known behavior: the swelling and deswelling of polymers based on their ionizable groups. Ionization makes these polymers water soluble when ionized, but they lose solubility in their neutral form. This shift is caused by a decrease in electrostatic repulsion as the ionizable groups neutralize, allowing hydrophobic interactions to dominate. The charge state, and thus solubility, can be reversibly altered by adjusting the pH of the surroundings⁹⁴.

pH-responsive enabled membranes

The pH-responsive membrane is made by integrating pH-responsive polymers into its structure. These membranes feature polyelectrolytes with ionizable acid or basic groups that can reversibly gain or lose protons depending on the surrounding pH. These changes lead to conformational adjustments in the membrane, causing the polymer chains to collapse or extend, which, in turn, alters the membrane’s surface properties and channel size⁹⁵. This section will explore recent advancements in pH-responsive smart membranes, focusing on dynamic pore size variation in response to pH, wettability, and stability of membranes over different pH ranges. Table 5 highlights recent research progress, although many challenges remain.

Table 5 pH-responsive membranes for different applications

Polystyrene-b-poly(4-vinylpyridine) (PS-b-P4VP)-based pH selective membranes possess the unique capability of adjusting their pore sizes in direct response to fluctuations in pH levels⁹⁶. This membrane reveals a tightly sealed pore configuration at lower pH levels, functioning with a permeance of roughly 10 L/m²/h/bar. In this acidic environment, the P4VP blocks lining the pore walls undergo protonation, triggering them to extend outward to negligible repulsion of charge, thereby effectively closing the pores. In contrast, under more alkaline conditions, P4VP chains are collapsed, which causes opening of PS-b-P4VP membrane pores, leading to a significant rise in permeance to over 800 L/m²/h/bar. This unique responsiveness to pH changes indicates the dynamic functionality of the membrane (Table 5, entry 1).

A group of researchers developed carbon nanotubes (CNTs) embedded with polyacrylic acid (PAA) intercalated between reduced graphene oxide (rGO), which enhances pH responsiveness and creates an additional pathway for water transport⁹⁷. In neutral and alkaline environment, the interaction between PAA and rGO improves separation efficiency, with a phosphate retention rate of 92.7% for 20 mg/L/bar (Table 5, entry 2) (Fig. 6A). The study also tested polyethylene glycol (PEG) 800 across various pH levels, finding that the membrane’s water permeability and phosphate rejection remained stable after five cycles from pH 3.0 to 8.2, confirming the durability and reversibility of its pH response. However, the lack of performance data under strongly alkaline conditions suggests a limited practical application due to the narrow pH response range.

Fig. 6: pH-responsive separation performance of the membrane.

A pH-responsive property of the RGO-gCNT membrane. Adapted with permission from ref. ⁹⁷. Copyrights (2019) Separation and Purification Technology. B pH-responsive ALG-modified membrane. Adapted with permission from ref. ⁹⁸. Copyrights (2019) American Chemical Society. C pH-responsive nanochannels of GO and GO-PEI membranes. Adapted with permission from ref. ¹⁰⁰. Copyrights (2021) American Chemical Society. D Schematic illustration of the pH-responsive performance of WSe2/PAA nano-gated membrane. Adapted with permission from ref. ¹⁰¹. Copyrights (2022) Chemical Engineering Journal. E Schematic illustration of pH-responsive GO/Gel membranes. Adapted with permission from ref. ¹⁰². Copyrights (2019) Journal of Membrane Science. Permission to reprint granted.

Using NIP technology, an NF membrane based on a five-block copolymer was developed with a pore size of 0.9 nm. Unlike the earlier example of carbon nanotubes (CNTs) grafted with poly(acrylic acid) (PAA), this copolymer’s pH-responsive conformational changes do not alter the retention of PEG 1000 or the membrane pore size (Table 5, entry 3) (Fig. 6B)⁹⁸.

This constancy was attributed to the high grafting density of the copolymer chains, which restricts the spacing between them. At 0.1 M ionic strength, the stability of water flux demonstrated that salt ions did not induce chain rearrangement. This reliable performance is particularly advantageous for size-based separations. Furthermore, the membrane exhibited a molecular weight cut-off (MWCO) of ~1 kDa, yet its permeability (13.0 ± 0.63–15.9 ± 0.06 L/m² h bar) exceeds that of commercial membranes with a MWCO of 1 kDa (15.4 L/m² h bar) and even surpasses that of some rated at 2 kDa⁹⁹. This advanced technology establishes a distinct cutoff value for molecular weight in nanofiltration membranes. If the pH response range (from pH 4–8.5) can be further refined, the application scope of these NF membranes will be significantly enhanced.

To tackle the balance between permeability and selectivity, positively charged PEI was grafted onto negatively charged GO nanosheets. This strategy mirrors the operational mechanism of glomeruli, which employs both molecular size screening and charge selectivity¹⁰⁰. The incorporation of the GO-PEI component significantly enhanced the hydrophilicity and surface charge of the membrane. Furthermore, this modification facilitated an increase in the size of the nano-channels due to the conformational changes induced by the PEI. Consequently, the pure water flux achieved a rate of 88.57 L/m²/h/bar, which is approximately four times higher than that of traditional GO membranes (Table 5, entry 4) (Fig. 6C). Membrane displayed pH-dependent rejection of Methylene blue (MB) and methyl orange (MO), at pH 12 and pH 2 levels, respectively, with a removal rate reaching as high as 96%.

As an alternative to GO nanosheets, a membrane with enhanced interlayer spacing between tungsten selenide (WSe₂) was developed by grafting polyacrylic acid (PAA) onto WSe₂ (Fig. 6D) (Table 5, entry 5)¹⁰¹. Such expansion enabled the membrane to tune the conventional trade-off between permeability and selectivity, achieving exceptional levels of both parameters. For example, at a pH of 3, the WSe₂/PAA membrane exhibited an average permeability of 1286.4 L/m²/h/bar and a solute rejection rate of 98.5%. Additionally, while the WSe₂/PAA membrane possesses a capillary structure analogous to the GO membrane, it features a reduced diffusion barrier for water molecules. Pressure-induced water transport reported good permeability and ion-selective capabilities of the membrane.

Inserting poly(N-isopropyl acrylamide-methacrylic acid) (PNIPAM-MAA) hydrogel into the graphene oxide (GO) layer allows for pH-responsive regulating nanochannels, simplifying the process by avoiding complex chemical grafting (Table 5, entry 6) (Fig. 6E)¹⁰². While hydrogels can deform in acidic environments (pH 2–6), GO layers’ stable size and consistent spacing create adjustable water channels that enhance small molecule separation compared to traditional NF membranes. The membranes’ responsiveness is determined by the combined effect of hydrogel volume fluctuations and the stable interlayer spacing of the GO sheets when exposed to variations in temperature and pH. They responded to both types of stimuli with high gating ratios and strong reversibility.

Another NF membrane has been developed utilizing the interfacial polymerization (IP) method, incorporating piperazine and chitosan¹⁰³. This membrane presents two distinct types of pores: a conventional polyamide (PA) network and chitosan beams. Notably, the common PA network pores maintain stability across varying pH levels, while the contraction of chitosan beams causes the expansion of chitosan beam pore sizes in alkaline conditions. Within the pH range of 2 to 6, the Donnan effect enhances the electronegativity of the membrane surface and increases the rejection rate of sodium sulfate (Na₂SO₄) from 36.8% to 97.7%. In the subsequent pH range of 6 to 10, the retention rate stabilizes at ~98%, attributed to the interplay between the Donnan effect and pore size screening (Table 5, entry 7). However, beyond a pH of 10, this equilibrium is disrupted, resulting in pore size screening predominating due to the membrane’s elevated electronegativity. The resulting expansion of the pores significantly enhances membrane flux; however, the rejection rate of Na₂SO₄ declines to 19.3%. Furthermore, dye rejection assessments indicate that in an alkaline environment, the pronounced electrostatic repulsion between anionic dyes and the negatively charged membrane surface counterbalances the impact of increased pore size. As a result, the membrane maintains high permeability while effectively repelling large dye molecules. Comparatively, the rejection percentage for small molecular dyes, such as methyl orange, remains suboptimal, achieving a rejection rate of only 10.0%. In conclusion, the membrane remarkably effectively separates dye and salt mixtures under alkaline conditions. This membrane demonstrates exceptional potential for effectively separating dye and salt mixtures under alkaline conditions. Further exploration into the mechanisms overriding the equilibrium between the Donnan effect and pore size screening is crucial to improve its effectiveness. Deepening this understanding will be key to advancing the design of smart membranes capable of dual-function regulation.

pH-responsive materials for membrane development and future directions

Recent advancements in research on pH-responsive membranes have led to the development of more innovative membrane materials. However, the established preparation technology for these membrane materials poses a significant challenge to their practical application in the future. While the current focus for new smart membranes emphasizes simplicity, sustainability, efficiency, and low cost, most of these innovations remain confined to laboratory experiments or small-scale simulations and tests. Overcoming the hurdles to commercialization remains difficult.

GO-based pH-responsive membranes exhibit low water permeability and poor antifouling performance due to the location of the -COOH groups. Addressing the trade-off between water flux and selectivity, especially for molecules with sizes close to the pore size. There is a lack of efficient methods for preparing liquid-dispersed WSe₂ nanosheets, which has limited research on 2D laminar WSe₂ separation membranes. The use of block copolymers may limit scalability and increase costs compared to traditional membrane materials. The PIP-CS/TMC membrane may have limited stability in strongly alkaline conditions (pH > 10), which could affect its practical application. Notably, recent work on 1 T′-MoS₂ laminate membranes has shown a breakthrough in overcoming some of these limitations. These membranes exhibit reversible and hysteretic water and ion transport modulated by pH, with a switching range of over two orders of magnitude in permeability. The hysteresis effect enables memory-like behavior, where the same pH value can produce permeating or non-permeating states depending on previous treatment. The mechanism is governed by interlayer ion exchange—particularly Li⁺ and H⁺—and dynamic interlayer spacing, demonstrating robustness and stability even over multiple cycles. This introduces a novel paradigm for intelligent membrane systems that combine responsive gating with molecular memory, enabling potential applications in biosensing, neuromorphic computing, and environmental monitoring¹⁰⁴.

Future initiatives should aim to enhance the stability, antifouling properties, and mechanical strength of smart membranes, enabling them to perform effectively in complex, real-world environments. Increasing their practical utility in terms of response range and speed will be crucial in overcoming the obstacles that hinder their commercialization and large-scale implementation. It is essential to further explore the development of block copolymer membranes that can operate reliably under various conditions while remaining cost-effective. Advancing the design of membranes with well-structured nanochannels and hierarchical architectures through the integration of diverse nanomaterials will greatly contribute to this field. Additionally, improving the compatibility and durability of smart membranes in extreme conditions, such as organic solvent nanofiltration, is imperative. Finally, by investigating the potential of 3D printing in manufacturing smart membranes, we can enhance resolution, material selection, and production efficiency, ultimately paving the way for a more innovative future.

Temperature-responsive materials (TRMs)

Temperature-responsive materials (TRMs), also known as thermo-responsive materials, are materials (e.g., polymers) that exhibit significant and abrupt changes in their physical or chemical properties by changing temperature. These changes result from temperature-induced alterations in molecular interactions, particularly in the balance between hydrophilic and hydrophobic forces. The ability to switch the features of molecules “on and off” enables a wide range of applications. These transitions are driven by changes in entropy, enthalpy, and hydrogen bonding^105,106. When heated, these polymers can transition from a hydrated, expanded state to a collapsed, hydrophobic state, often accompanied by significant changes in solubility. These polymers undergo phase transitions at certain temperatures, known as Lower Critical Solution Temperature (LCST) or Upper Critical Solution Temperature (UCST) behavior^107,108. Based on the critical temperature, polymers are classified into LCST and UCST polymers, as shown in Table 6^{109,110,111,112,113,114,115,116,117,118}.

Table 6 Typical thermo-responsive materials, their structure, and their LCST/UCST temperature

Temperature-responsive polymers (TRPs)

Lower critical solution temperature (LCST) polymers

LCST-type TRPs exhibit solubility in aqueous or organic solvents at temperatures below their lower critical solution temperature (LCST). Once the temperature exceeds this threshold, these polymers undergo a notable phase transition and become hydrophobic due to enhanced intra- and intermolecular hydrophobic interactions¹¹⁹. They change their conformation from coil to a globular form ref. ¹²⁰. This transition typically results in phase separation, aggregation, or reduced swelling¹²¹. At higher temperatures, the breaking of H-bonds creates thermodynamically favorable interactions between polymer and solvent molecules. As a result, polymer chains become more thermodynamically stable. This phenomenon occurs when the negative enthalpy of mixing (ΔHmix) and entropy of mixing (ΔSmix) are balanced due to increased molecular ordering¹²². Thus, LCST behavior is primarily determined by entropy. Polymers that display LCST behavior in water include polyethers, poly(vinyl alcohol) (PVA), poly(N-alkyl-substituted acrylamides), poly(methyl vinyl ether), poly(2-oxazoline), poly(vinyl caprolactone) (PVCL), and modified agarose and cellulose^123,124,125. These polymers undergo rapid and reversible phase transitions triggered by temperature changes¹²⁶. Among thermo-responsive polymers, poly(N-isopropylacrylamide) (PNIPAAm) and its derivatives are the most widely studied. PNIPAAm exhibits temperature-dependent reversible coil to globular conformation in water at its LCST of 32 °C. This LCST is very close to human body temperature, making it particularly appealing for biomedical applications. Furthermore, the LCST can be precisely tuned by introducing substituents such as a carboxyl group (2-carboxyisopropylacrylamide, CIPAAm)¹²⁷, an amine group (2-aminoisopropylacrylamide, AIPAAm)¹²⁸, or a hydroxyl group (2-hydroxyisopropylacrylamide, HIPAAm)¹²⁹, as well as through copolymerization with functional monomers. Such modifications enable these polymers to respond to various PH, salt concentration, temperature and ionic strength.

Upper critical solution temperature (UCST) polymers

UCST polymers behave differently from LCST polymers; they exhibit phase separation when cooled down. LCST-type polymer solutions exhibit a homogeneous and transparent state when the temperature falls below the lower critical solution temperature (LCST). In contrast, UCST-type polymer solutions display a heterogeneous, cloudy, and opaque appearance at temperatures beneath the upper critical solution temperature (UCST)^130,131. Unfortunately, there are relatively few well-established examples of UCST-type thermo-responsive polymers. Moreover, their phase transitions are typically less pronounced, and their temperature ranges are less sharp than those of LCST polymers. For instance, interpenetrating networks of polyacrylamide (PAAm) and polyacrylic acid (PAAc) exhibit a UCST at ~25 °C. At temperatures below the UCST, AAm and AAc form stable hydrogen bonds, but when the temperature rises above the UCST, these bonds break, leading to network swelling^123,124,132. Ureido-based polymers, including poly(allylurea) (PU) and poly(l-citrulline) derivatives, have demonstrated UCST behavior in buffer solutions^133,134. The UCST of these systems can be controlled by adjusting the ureido ratio and molecular weight. Additionally, a series of poly(allylurea-co-allylamine) (PU-Am) copolymers have shown adjustable UCST values ranging from 8 to 65 °C by adjusting the amino group content¹³⁵.

Temperature-responsive enabled membranes

Temperature-responsive membranes are designed by integrating thermally sensitive polymers that undergo reversible conformational changes at specific temperature thresholds. These membranes primarily utilize polymers with LCST or UCST behavior, which undergo dramatic hydrophilic-hydrophobic transitions by changing the temperature. These polymers switch between extended and collapsed states upon temperature changes, altering the membrane’s pore size, permeability, and surface properties^18,136. This section explores recent advancements in temperature-responsive membranes, focusing on controlled molecular transport, switchable surface properties, and performance across varying temperature ranges.

The development of highly hydrophilic membranes with excellent separation capabilities remains a challenge. PNIPAM-grafted-Dextran was used to modify polyethersulfone (PES) membranes, enhancing their functionality and performance (Table 7, entry 1)¹³⁷. Incorporating the dextran segment, which is rich in -OH groups, imparted the modified membrane with superior hydrophilicity and antifouling properties. This improvement is reflected in the reduction of the WCA on the membrane surface from 89.3° to 58.8° compared to the original PES membrane. As the temperature increased from 25 °C to 40 °C, PNIPAM chains contracted, enlarging the pore size and increasing pure water flux, but also reducing BSA rejection due to pore blockage. Achieving high antifouling properties requires precise adjustment of the PNIPAM ratio. However, the membrane is ineffective for separating smaller dye molecules, such as MO, MB, and rhodamine B, with near-zero rejection rates, which limits its use for small-sized substances.

Table 7 Comparison of different temperature-responsive membranes with their operational temperature, membrane fabrication, pore size change during temperature change, and practical application

Adding thermo-responsive nanofillers is an effective way to introduce temperature sensitivity into thin-film nanocomposite (TFN) polyamide NF membranes. In this approach, thermo-responsive zwitterionic nanocapsules (ZNCs) are incorporated within the polyamide (PA) active layer to establish temperature-sensitive water transport channels, improving nanofiltration performance (Table 7, entry 2)¹³⁸. When the temperature exceeds the UCST, the zwitterionic polymer chains fully extend, leading to an abrupt increase in the size of the zinc nanotube channels and facilitating faster water transport (Fig. 7A). The permeability of the ZNCs-TFN membrane increased from 57 L/m²/h to 88 L/m²/h at 25 °C and 30 °C, respectively, while maintaining its salt rejection capability. Despite considerable research efforts, fouling remains the primary obstacle for nanofiltration membranes, leading to extra transport resistance and necessitating higher operating pressures in real-world use. To address this, a highly antifouling nanofiltration membrane was created by grafting poly(N-isopropylacrylamide) chains onto a brominated polyamide layer (PA-Br), resulting in twice the permeance of the pristine membrane (PA) while maintaining effective multivalent ion rejection (Table 7, entry 3)¹³⁹.

Fig. 7: Temperature-responsive separation characteristics of the membrane.

A Schematic representation of the thermal-responsive behavior exhibited by ZNCs-TFN membranes. Adapted with permission from ref. ¹³⁸, Copyrights (2022) Journal of Membrane Science. B Normalized flux decline curves for the PNIPAM-grafted thin-film composite (TFC) membranes were evaluated during three distinct filtration periods using BSA solution. C, D illustrate the schematic representation of the antifouling mechanisms associated with PNIPAM-grafted TFC membranes. Adapted with permission from ref. ¹³⁹. Copyrights (2022) American Chemical Society. E Schematic picture of switchable wettability of the PNIPAAm decorated membrane driven by heating¹⁴⁴. Permission to reprint granted.

In addition, PNIPAM chains displayed excellent antifouling properties during both filtration and traditional cleaning stages. Furthermore, at a water temperature above LCST, the PNIPAM chains shrank, establishing a buffer layer to separate the membrane from pollutants (Fig. 7B–D). The PNIPAM/PA-Br membrane retained its ability to reject multi-valent salts and achieved a high flux of 18.2 L/m²/h, nearly double that of the original PA membrane. These results establish a foundation for the synthesis of advanced, stimuli-responsive polymeric chains designed to create antifouling membranes.

Most polymer-based stimuli-responsive membranes typically exhibit unidirectional gating behavior, either negative or positive. Developing membranes with reversible gating has become an important research focus. In one approach, PNIPAM was grafted onto GO membranes by simply tuning the molecular grafting density (Table 7, entry 4)¹⁴⁰). The permeability of water and small molecules through the membrane can be regulated by adjusting the surrounding temperature. The reversible thermal response gating mechanism is driven by conformational changes in the PNIPAM grafted layer, which depend on the grafting density of PNIPAM on the membrane surface.

The conformational changes of PNIPAM chains at different grafting densities drive the reversible temperature-responsive gating mechanism. At lower grafting densities, PNIPAM chains began to form horizontal aggregates on the GO surface. When the temperature exceeds the LCST, the shrinkage of PNIPAM chains creates more space for water to pass through, resulting in positive gating characteristics. At higher grafting densities, steric hindrance between PNIPAM chains induces conformational changes in the upright direction. When the temperature exceeds the LCST, PNIPAM chains form intermolecular or intramolecular hydrogen bonds, causing the polymer chains to shrink. Intermolecular hydrogen bonding decreases the channel size, resulting in negative gating characteristics. In contrast, intramolecular hydrogen bonding increases the nanochannel size, producing positive gating characteristics.

Membranes were fabricated via immersion phase inversion technique using a casting solution containing polyvinylidene fluoride-g-poly(N-isopropylacrylamide) (PVDF-g-PNIPAM) and graphene oxide (GO) nanosheets (Table 7, entry 5)¹⁴¹. Moreover, upon illumination, the membrane rapidly heats up, causing the contraction of PNIPAM chains and the consequent expansion of the membrane pores. By integrating this reversible, light-triggered structural change, the membrane demonstrated a high-water gating ratio when the light was switched on and off. With just 0.095 wt% of GO, the membrane could reach a temperature of 56 °C within 200 seconds under illumination in a water flow state. The permeability of the membrane was sensitive to environmental changes; notably, its water gating ratio reached as high as 1.72 when the liquid temperature increased from 25 °C to 40 °C.

Similarly, membranes of the same type were prepared by blending PVDF-g-PNIPAM polymer and graphene oxide (GO) with PVDF to produce temperature-responsive separation membranes (Table 7, entry 6)¹⁴². The results demonstrated that varying the NIPAM-to-PVDF ratio enabled precise control over the NIPAM grafting levels in the polymers. As the temperature increased from 22 °C to 40 °C, the membranes exhibited decreasing rejection rates. This behavior is due to the contraction of PNIPAM side chains at higher temperatures, leading to large membrane pore sizes and reduced rejection performance. In another study, commercial poly(ethylene terephthalate) (PET) microfiltration (MF) membranes were functionalized with thermoresponsive PNIPAM using a polydopamine coating (Table 7, entry 7)¹⁴³. Initially, dopamine was self-polymerized under aqueous conditions to form a reactive layer. Subsequently, amino-terminated PNIPAM was grafted onto this polydopamine layer via its amino groups. The pure water flux nearly doubled when the temperature was raised from 20 °C to 45 °C. In contrast, the water flux of the unmodified PET membrane remained unchanged over this temperature range, as its structure was unaffected by heating. The membrane separation performance data show that the membrane effectively rejected the protein BSA, achieving over 80% rejection.

A thermoresponsive Poly(N-isopropyl acrylamide) (PNIPAAm)-modified nylon membrane with switchable wettability was fabricated using a simple hydrothermal method (Table 7, entry 8)¹⁴⁴. At a temperature of 25 °C (below LCST), the membrane is hydrophilic (~20°) whereas at a temperature of ~45 °C (above LCST), the membrane is super hydrophobic with a water contact angle of ~120 °C (Fig. 7E). The membrane demonstrates strong potential for practical applications, including on-demand oil spill cleanup, fuel purification, and industrial wastewater treatment.

Developing functional membrane systems is vital for achieving high-performance functionality. Two-dimensional titanium carbide (MXene) was used as a model substrate, and poly(N-isopropyl acrylamide) was endowed with dual-level regulatable nanochannels, achieved through adjustments to the nanochannel microenvironments (Table 7, entry 9)¹⁴⁵. This hybrid nanochannel regulation strategy integrates dynamic thermo-responsive behavior with permanent structural reconfiguration to construct high-performance MXene-based separation membranes. The stimulus-responsive component is enabled by poly(N-isopropylacrylamide) (PNIPAM), whose reversible conformational transition across its lower critical solution temperature (LCST) allows real-time tuning of nanochannel size, imparting controllable permeability and molecular selectivity. In parallel, the intrinsic nanochannel framework is fine-tuned through in situ oxidation of Ti₃C₂Tₓ MXene nanosheets, resulting in the formation of TiO₂ and amorphous carbon. This irreversible transformation alters the chemical microenvironment, interlayer interactions, and nanochannel geometry, thereby enhancing membrane compactness and baseline selectivity. Although this oxidation process is not reversible and thus does not constitute a responsive feature, it significantly influences the behavior of the PNIPAM layer by modulating its binding affinity and spatial distribution within the membrane matrix. The combination of these two regulatory mechanisms, reversible, temperature-driven PNIPAM gating and irreversible, oxidation-induced nanochannel compaction, enables precise, gradient molecular separation across a wide size range.

Temperature-responsive materials for membrane development and future directions

Temperature-responsive membranes have gained significant attention due to their ability to adapt to varying thermal conditions, offering dynamic control over permeability and selectivity. Various designs of TRPs have been employed to optimize their performance. Surface modification using PNIPAM (poly(N-isopropylacrylamide)) is widely employed in developing temperature-responsive membranes, with particular attention given to its functional group modifications. The chemical structure of PNIPAM can be tailored by introducing different functional groups to adjust its LCST behavior and membrane performance. For instance, incorporating hydrophobic groups like n-butyl methacrylate can lower the LCST, while adding hydrophilic groups such as acrylic acid increases it¹²⁷. These modifications not only affect the temperature response range but also influence the switching kinetics and reversibility of the membrane. Incorporating various functional groups can also enhance specific properties. Adding charged groups can improve protein resistance, while introducing cross-linkable groups can enhance the stability of the surface modification. The strategic modification of PNIPAM’s chemical structure through different functional groups thus provides a versatile approach to fine-tune membrane performance for specific applications

The density of the functional groups is a crucial strategy for enhancing the performance of temperature-responsive membranes by varying the densities of temperature-responsive polymer within the membrane matrix. Higher concentrations of temperature-responsive polymer units can lead to increased responsiveness and faster switching times, resulting in more pronounced changes in permeability upon temperature¹⁴⁶. The density of functional groups can also influence the membrane microstructure and phase behavior, allowing for tailored permeability profiles and dynamic gating mechanisms¹⁴⁷.

In addition to thermoresponsive polymers, two-dimensional materials such as titanium carbide (MXene) and graphene oxide have been integrated into membrane designs to enhance their performance. These materials possess intrinsic properties like high surface area, mechanical strength, and the ability to form responsive nanochannels. MXene, for instance, can undergo spontaneous oxidation, which can alter the size of nanochannels, providing an additional level of regulation for the membrane’s functionality. By integrating 2D materials with thermoresponsive polymers, a dual-regulation system is established, where the polymer’s phase transition and the material’s intrinsic properties collaborate to control the behavior of the membrane dynamically.

While these advancements have demonstrated the potential of temperature-responsive membranes in various applications, several challenges remain. For example, controlling the precise temperature-responsive behavior and ensuring the stability and scalability of these membranes for industrial applications is still a work in progress. Furthermore, fouling and long-term performance degradation must be addressed to improve their longevity and practicality in complex applications like wastewater treatment, bio-separation, and drug delivery systems. While significant strides have been made in developing temperature-responsive membranes, further research is needed to refine these systems and unlock their full potential for a wider range of practical applications.

Other responsive mechanisms

Along with the four types of stimuli-responsive membranes mentioned earlier, other response mechanisms, including salt responsiveness and magnetic-responsiveness, have also been documented in the literature. This section will provide a brief overview of both salt-responsive and magnetic-responsive membranes.

Salt-responsive membranes

The power of tailored membrane design can be harnessed to enhance NF membrane separation performance under different salt concentrations. By adjusting pore size and electrical properties, remarkable results can be achieved. High pore size uniformity is within reach through controlled diffusion of reaction monomers and the incorporation of nanoparticles. Additionally, methods such as surface coating, grafting, and the incorporation of nanomaterials, along with new monomers, have been shown to modify the electrical characteristics of membranes. This includes influencing the membrane’s charge and dielectric constant.

The poly(amide) (PA) nanofiltration (NF) membrane incorporating zwitterionic nanospheres (PA-ZNs) with salt-responsive ion valves boosted salt rejection and improved selectivity between monovalent and divalent ions even at high salt concentrations, overcoming the challenges of treating brackish water with elevated salinity¹⁴⁸. The ion valve effect, caused by the “confinement-filling” interaction between the nanospheres and the polyamide layer, was critical in regulating the membrane separation performance at high salt concentrations. The main separation mechanisms utilized by the PA-ZNs membrane were steric exclusion and the dielectric effect. The rejection rate of the PA-ZNs slightly dropped to 96.2% when the Na₂SO₄ concentration rose to 5 g/L, but then it consistently improved, reaching 98.3% at a higher salinity of 9 g/L.

Salt-responsive polyamide thin-film nanocomposite (TFN-ZNP) membrane incorporating zwitterionic polymeric nanoparticles (ZNPs) was constructed for high-efficiency dye desalination¹⁴⁹. The salt-responsive behaviour of the ZNPs enabled the TFN-ZNP membrane to adjust its microstructure and separation performance in response to changes in NaCl concentration in the feed solution. The TFN-ZNPs membranes exhibited increased selectivity for dye/dye and salt/dye separations compared to the control polyamide membrane. The water permeance of the TFN-ZNPs increased to 1.7 times that of its pure water permeance when tested with a 10 g/L NaCl solution, demonstrating a reversible salt-responsive behavior. Additionally, this salt-responsive behavior gave the membrane excellent separation ability for both salt/dye and dye/dye mixtures. The selectivity of the TFN-ZNPs improved significantly, reaching around 104 for NaCl/Congo red and 127 for methyl orange/neutral blue pairs. Furthermore, the TFN-ZNPs demonstrated enhanced antifouling performance, with their flux recovery rate rising from 85.6% to 96.5% when cleaned using a NaCl solution rather than deionized water. Polyacrylic acid (PAA) chains on the surface of membranes provide numerous reaction sites for zwitterionic nanogels, which attach to these chains and form structures resembling clusters of grapes¹⁵⁰. In salt solutions, the PAA chains curl into coils and move in synchrony with the amphoteric nanogels, enhancing both the effective pore size and the density of responsive components. This coordinated movement effectively addresses the limitations of traditional stimuli-responsive membranes, which often exhibit slow response times and restricted flexibility. An ultrasensitive salt-responsive membrane was fabricated by grafting zwitterionic nanohydrogels onto a PAA-g-PVDF microporous membrane, which exhibited an ultrahigh gating ratio of up to 89.6 times for Mg²⁺ ions and a very rapid response to changes in salt concentration.

Salt-responsive membranes are gaining attention for their ability to adapt to temperature changes, allowing dynamic control over permeability and selectivity. One key approach involves modifying poly(N-isopropylacrylamide) (PNIPAM) to enhance membrane performance by adjusting its functional groups and Lower Critical Solution Temperature (LCST) behavior.

Thermo-responsive and salt-responsive UF membranes with antifouling capabilities can significantly improve water treatment efficiency. Preliminary tests have shown that a PNIPAM-co-PDMAC copolymer, which responds to both temperature and ionic strength, can reduce irreversible fouling and enhance membrane durability through hydration transitions triggered by salinity changes. The copolymer’s hydration state and dynamic pore size modulation are sensitive to changes in salinity and temperature, with salt-triggered hydration transitions reducing irreversible fouling and extending membrane durability. NF membranes with salt-responsive ion valves could expand applications in brackish water treatment.

Future research should focus on innovative materials and fabrication strategies. Modified and hybrid Membrane Bioreactor (MBR) systems are recommended for higher salinities, leveraging halophilic bacteria for hypersaline wastewater treatment. Additionally, novel 2D materials-based membranes show promise in desalination by effectively excluding small ions. Size exclusion based on sub-nanometer pores facilitates the rejection of small ions like sodium at high hydraulic pressures and relevant feed salt concentrations. Molecular layer-by-layer assembled thin-film composite membranes have demonstrated improved salt rejection and water flux, along with better antifouling performance than traditional reverse osmosis membranes.

Magnetic-responsive membranes

Magnetic-responsive membranes consist of superparamagnetic nanoparticles attached to polymer chains on their surfaces. When exposed to oscillating magnetic fields, the nanoparticles cause movement or heating of the polymer chains, improving membrane performance. This method reduces concentration polarization and fouling while enabling remote control of molecular sieving properties. Various combinations of polymers and nanoparticles, such as poly(2-hydroxyethyl methacrylate) PHEMA with magnetite and PNIPAm with iron oxide, have been studied^{151,152,153,154}. These magnetically responsive membranes significantly enhance flux, rejection rates, and adjustable molecular weight cutoffs, making them promising for bioseparation and water treatment applications.

Magnetically responsive UF membranes were developed by grafting magnetically doped PHEMA chains onto the top of cellulose UF membranes. Movement of magnetic nanoparticles at the PHEMA chain ends induced movement in the grafted polymer chains, which led to a decrease in permeate flux from 32 L/m²/h to 17 L/m²/h for BSA and 206 L/m²/h to 31 L/m²/h for dextran, but an increase in rejection for both BSA (90%) and dextran (40%).

A magnetoresponsive mesoporous membrane was developed that allows for remote-controlled, stepwise tunable adjustments of sieving barrier pore sizes. This innovation was achieved by incorporating iron oxide nanoparticles into a thermo-responsive block copolymer matrix¹⁵³. The membrane’s pore size can be tuned by applying an alternating magnetic field (AMF), which causes the heating of embedded iron oxide nanoparticles and triggers conformational changes in the thermo-responsive polymer on the pore walls. When exposed to an alternating magnetic field with varying input energy, the water permeability and dextran rejection of the thin-film mixed matrix nanocomposite membranes were adjusted to different levels. This indicates that the pore size distribution of the sieving barrier in these membranes can be adjusted in a stepwise manner.

Despite the potential of magnetic-responsive membranes, challenges remain in optimizing membrane performance and expanding their applications. Future research is needed to address the issues related to low magnetic responsiveness, difficulty controlling magnetic actuation, and the potential cytotoxicity of magnetic nanomaterials.

Magnetic-responsive materials for membrane development and future directions

Despite the potential of magnetic-responsive membranes, challenges remain in optimizing membrane performance and expanding their applications. Future research is needed to address the issues related to low magnetic responsiveness, difficulty controlling magnetic actuation, and the potential cytotoxicity of magnetic nanomaterials.

Magneto-responsive membranes can be created by combining organic polymer-based membranes with inorganic magnetic nanoparticles, allowing for switchable separation properties to achieve reversible changes in barrier and surface properties, which is a promising direction for future applications in bioseparations and biomedical fields. This could be done through two main approaches: addressing secondary interactions or focusing on intrinsic membrane barrier properties.

The development of anisotropic magnetic-responsive membranes is another interesting research area for achieving quick magnetic responses. This could be achieved by incorporating magnetic particles into an elastic polymeric matrix, either randomly or in an ordered structure. A uniform magnetic field applied during cross-linking causes particle chains to form and lock into the elastomer, leading to anisotropic properties. The magnetic particles couple the shape and elastic modulus with the external magnetic field, resulting in giant deformational effects, high elasticity, and anisotropic elastic and swelling properties.

General conclusions & future outlooks

This review presents an inclusive analysis of advancements in stimuli-responsive membrane development. The chemistry, response mechanisms, and applications of stimuli-responsive materials were discussed and evaluated in terms of performance and, whenever possible, scalability. These innovative membrane materials, which exhibit active responses, hold great promise in the fields of environmental science and energy management. Most stimuli-responsive membranes are designed to respond to external stimuli by incorporating stimuli-responsive components into the membrane materials. Unlike conventional nanofiltration membranes, these smart membranes can dynamically adjust their properties in response to external stimuli, enabling effective separation even in complex environments. Additionally, their advantages, such as low manufacturing costs, reduced energy requirements, and broad responsiveness, facilitate easy operation in real-world scenarios.

Stimuli-responsive membranes, however, remain largely in the realm of basic research, with most studies limited to proof-of-concept demonstrations. While the potential of these systems is clear, further research is necessary to overcome the key technical challenges outlined in this review. At the same time, there is a growing need to shift focus toward the development of practically relevant, application-driven technologies that leverage these smart membranes. Overcoming the challenges posed by emerging pollutants and developing new responsive materials that allow more precise control of membrane properties through adjustments in the chemistry of responsive materials and different membrane fabrication strategies. Demonstrating such real-world use cases will be crucial in the coming decades to bridge the gap between lab-scale innovation and scalable solutions.

The energy input into photo-responsive, electro-responsive, salt-responsive, and magnetic-responsive polymers is crucial for enhancing their functionality and expanding their applications across various fields. In photo-responsive polymers, light energy effectively triggers conformational changes, resulting in improved physical properties. For instance, exposure to UV light can cause the polymer chains to expand or contract, leading to innovative uses in optical devices, responsive sensors, and advanced drug delivery systems. Electro-responsive polymers respond to electrical stimuli, significantly influencing their conductivity and mechanical properties. This capability enables exciting applications in soft actuators and artificial muscles, where precise control and adaptability are essential. Salt-responsive polymers can modify their behavior based on the ionic strength of the surrounding environment, especially in terms of swelling and mechanical characteristics. This responsiveness is particularly beneficial in drug release systems, as it can regulate the release rates of active ingredients, optimizing therapeutic effects. For magnetic-responsive polymers, utilizing magnetic fields can induce movement or reorientation. This ability is valuable for targeted drug delivery systems and innovative actuation technologies, allowing for high levels of control over the polymer’s behavior.

Combining different stimuli can finely tune the properties of these polymers. For example, spiropyran derivatives exhibit both photo-responsiveness and pH-responsiveness, enabling them to react quickly and reversibly to changes in light and acid/base concentration. These derivatives can function as multi-responsive single-molecule electrical switches in nanoscale devices. Their electrical switching mechanism is based on the difference in conductance between two states: the closed form, which has broken conjugation and low conductivity, and the open form, which has complete conjugation and high conductivity. This unique property allows them to respond to both light and acid stimuli. Such multi-responsive systems are highly versatile with promising applications in soft robotics, advanced drug delivery, and the design of organic circuits.

In summary, the influence of energy input on these responsive polymers not only enhances their properties but also opens up new and innovative applications in biomedical engineering, materials science, and environmental solutions. By harnessing these interactions, we can develop cutting-edge technologies that address diverse needs and challenges.