Molecularly engineered cellulose: the next-generation sustainable polymer electrolyte material

Abstract

Growing environmental imperatives are driving the need to substitute petroleum-derived materials with renewable and sustainable alternatives to enable the production of biodegradable and carbon-neutral products. As a naturally abundant and versatile biopolymer, cellulose has been extensively utilized in conventional industries such as papermaking and textiles and is increasingly being applied in emerging advanced fields, including energy storage, food technology, emulsions, coatings, cosmetics, and biomedical applications. With the iteration and development of energy technology, cellulose-mediated polymer electrolyte materials (PEMs) have re-emerged as the materials of interest to notable scientific and commercial communities due to their exceptional performance advantages in electrochemical energy storage. In this review, we comprehensively summarize and analyze the molecular engineering strategies, key features, and the corresponding construction strategies utilizing cellulose for the preparation of novel PEMs. Particularly, we provide a material and structural perspective on how the ionic conductivity, ion selectivity, anti-swelling properties, self-healing properties, flame retardancy, porosity, mechanical properties, and photoelectric stability of cellulose-mediated PEMs can be regulated through molecular chemistry. Finally, we examine the potential of these strategies in advancing circular economy principles and environmental sustainability objectives, while also identifying key challenges and outlining promising future research directions. We emphasize the critical need for advanced molecular-level chemical engineering to fully harness the potential of cellulose for energy-related applications.

---

Broader context

Increasing environmental concerns require the replacement of petroleum with renewable and sustainable resources to produce biodegradable and carbon-neutral products. As a natural biopolymer, cellulose has long been used in traditional fields such as papermaking and textiles and is now gradually emerging in advanced fields such as energy storage, food technology, emulsions, paints, cosmetics, and healthcare. With the continuous iteration and development of energy technology, cellulose-mediated polymer electrolyte materials (PEMs) have re-emerged as the materials of interest to notable scientific and commercial communities due to their exceptional advantages in terms of preparation cost, structural design possibilities, and material performance. In this regard, exciting research activities ranging from fabrication strategies to energy-related applications of cellulose-mediated PEMs have emerged, and their future is even more exciting. This review redefined the construction strategy and design principles of cellulose-mediated PEMs from a molecular engineering perspective, comprehensively described the most prominent application scenarios for cellulose-mediated PEMs in recent years, and provided a clear and detailed interpretation of the required properties for cellulose-mediated PEMs and the common design approaches across different applications. This review is a timely addition to cellulose-mediated PEMs for energy-related applications and is expected to further promote the research enthusiasm of cellulose materials in energy-related applications and provide comprehensive and detailed references for the design and application strategies of state-of-the-art cellulose functional materials focusing on energy devices.

1. Introduction

The escalating issues of environmental pollution and global energy consumption, amplified by the continued use of fossil fuels, are driving a worldwide transition toward the development of renewable, sustainable, and environmentally benign energy solutions.^1,2 In particular, significant strides have been made in the development of highly efficient and advanced energy storage technologies designed to convert sustainable energy into transportable and stable chemical energy forms. In this regard, a diverse range of energy storage devices, featuring various design configurations (e.g., prismatic, cylindrical, coin-type, magazine, or blade stacks) and specialized functionalities, have been developed.^3,4 In parallel, novel battery chemistries have been proposed to meet the stringent requirements of extensive application scenarios, ranging from portable consumer electronics and electric vehicles to large-scale power storage systems for smart grids. Indeed, a major avenue for advancing these energy storage devices is the discovery of key electrode materials or PEMs with higher performance, lower cost, and enhanced sustainability.^5–10

In recent decades, polymer electrolyte materials (PEMs) have re-emerged as the materials of interest to notable commercial and scientific communities for electrochemical energy storage.^11,12 This interest was initially stimulated by the high modulus of solid ionic conductors, which were thought to provide the most direct route to practical, high-energy, and rechargeable batteries based on metal anodes.¹³ Recent interest in PEMs is multifaceted, propelled by several concurrent trends: the advent of novel high-performance materials, growing safety imperatives, and the expansion into new application domains.¹⁴ In this regard, it is notable that certain PEMs with different chemical properties can achieve room temperature ionic conductivity (∼1 mS cm⁻¹) comparable to that of a conventional liquid electrolyte. Moreover, the flexibility of PEMs facilitates contact with the electrodes, resulting in better electrode/electrolyte interfacial properties or achieving higher durability. In general, most PEMs are based on crystalline polymers, and the ionic conduction occurs mainly in the non-crystalline regions of the polymer. Taking poly(ethylene oxide) (PEO) as an example, the dissociated cation can coordinate with the O atoms of the ethylene oxide group on the PEO backbone and migrate along the chain segments under the influence of an electric field. Consequently, the movement of the polymer chain segments of PEMs directly affects ionic conduction, and the ionic conduction is generally accompanied by the formation and breakage of ligand bonds. In brief, the ions are hopping both within and between the chains of the polymer, coupled with the segmented movement of the polymer chains, which ultimately allows the polymer to conduct ions. Indeed, besides the mentioned properties such as ionic conductivity and flexibility, practical factors including environmental sustainability, cost-effectiveness, and processability of PEMs must be taken into account to assure the long-term commercial viability of these next-generation PEMs.^15–20

Cellulose, as the most abundant renewable biopolymer on Earth with an annual production of around 1.5 × 10¹² tons, represents a highly promising alternative to petroleum-based materials due to its inherent biodegradability and carbon-neutral lifecycle.^21–25 Constituting approximately 50% of cell walls of terrestrial plants, cellulose can be extracted from plant biomass and processed or functionally tailored into high-performance materials, offering enhanced sustainability and reduced environmental impact.^26–28 The abundance of hydroxyl groups on the cellulose skeleton offers a dynamic platform for its precise molecular and structural design, which has given rise to a variety of industrially produced cellulose derivatives (e.g., cellulose esters and cellulose ethers) with continuous expansion of its use into research domains beyond its traditional scope.^29–33 Depending on different treatments, cellulose can also be further obtained using nanoscale and special property materials, such as cellulose nanofibers, cellulose nanocrystals, and bacterial cellulose. Noticeably, research on the conversion of cellulose into high-performance PEMs through molecular engineering strategies has gained significant and rapid development progress in recent years.^34,35 In this regard, depending on the treatment strategies of chemical modification, ionic/molecular cross-linking, physical blending or dissolution regeneration, and the preparation methods of solution recasting, vacuum-filtration, electrostatic spinning or phase conversion, cellulose is masterfully prepared into a new-generation of PEMs, which have demonstrated advanced and unique properties in a wide range of energy storage devices (Fig. 1a and b). Over the last decade, the focus of cellulose-mediated PEM research has expanded from basic physical blending to novel chemical treatments to enable innovative applications in energy storage.

Fig. 1 General overview of cellulose as a PEM. (a) Schematic diagram of the characteristics, key requirements, and applications of cellulose-mediated PEMs; (b) The number of publications about cellulose-mediated PEMs according to the Web of Science; (c) Radar plots of liquid electrolyte materials, traditional polymer electrolyte materials, and cellulose-mediated electrolyte materials (the trend originates from ref. 36. Copyright 2024, Elsevier).

Although several reviews of cellulose-based separators or cellulose-based solid polymer electrolytes have been reported, there is an urgent need to comprehensively summarize and analyze the advances in cellulose-mediated PEMs from new perspectives, given the rapid iterations and developments in this field. Therefore, in this review, we provide the first comprehensive overview of multiscale engineering strategies for converting cellulose into novel PEMs through molecular engineering, including chemical modification (esterification/oxidation/sulfonation/phosphorization/etherification reaction), cross-linking (ionic/molecular), molecule entanglement (immersion curing/molecule bonding), and dissolution regeneration (ionic liquid/alkali–urea/LiBr–formic acid/AlCl₃–ZnCl₂/NaOH). Moreover, we also summarize and discuss in depth the key features achievable in cellulose-mediated PEMs and the corresponding construction strategies, aiming to provide a complete and clear view of the properties of cellulose-mediated PEMs (Fig. 1c and Table S1). Meanwhile, the recent developments in cellulose-mediated PEMs for rechargeable lithium batteries, rechargeable zinc batteries, rechargeable sodium batteries, rechargeable aluminum batteries, supercapacitors, fuel cells, reverse electrodialysis systems, and solar cells are highlighted. It is sincerely expected that this review will provide some inspiration for the design and application of cellulose-mediated PEMs in the future.

2. Opportunities for constructing PEMs using cellulose

2.1 Key requirements and characteristics of PEMs

The substantial research focus on PEMs has been predominantly motivated by their improved safety profiles and compatibility with battery systems requiring high energy density and mechanical flexibility.^37–39 Indeed, the safety of energy-related devices is the top priority, especially for high-energy storage systems. Generally, liquid electrolytes have high ionic conductivity and good interfacial contact with electrodes, but lower ionic transfer coefficients, higher leakage, and reactivity severely limit their development. Specifically, the exothermic reaction that occurs between the electrodes and electrolyte causes the overall temperature of the battery to rise rapidly, triggering further decomposition of the electrolyte and the generation of flammable gases.^40,41 In comparison, PEMs have excellent (electro)chemical stability, which contributes to reducing or mitigating the heat generated by the battery in the event of thermal shock or thermal runaway.^42–44 Moreover, in terms of basic physicochemical properties, PEMs offer multiple advantages, including straightforward synthesis, low cost, superior chemical stability, low mass densities, compatibility with large-scale manufacturing processes, and the inherent mechanical toughness at temperatures well above the glass transition temperature. In general, PEMs are composed of a polymer matrix with alkali metal salts dissolved in the polymer matrix. During about 50 years of development, a variety of polymer materials have been employed to prepare novel PEMs, such as poly(ethylene oxide) (PEO), poly(methyl methacrylate) (PMMA), poly(acrylonitrile) (PAN), poly(ethylene succinate) (PES), poly(vinyl alcohol) (PVA), polyethylene (PE), poly(vinylidene fluoride) (PVDF), poly(ethylene glycol) methyl ether methacrylate (PEGMA), poly(ethyl cyanoacrylate) (PECA), and poly(propylene oxide) cyclocarbonate terminated (PPC) (Fig. 2a).

Fig. 2 Key properties and the corresponding construction strategies of PEMs. (a) Common polymer electrolyte materials (PEMs). (b) Achievable properties of PEMs and the corresponding construction strategies.

For PEMs, the polymer chain movement, the dissociation energy of ions leaving the polymer chains, and ion transport pathways are key factors that drive four strategies for enhancing the electrochemical properties of PEMs: (i) reduce the crystallinity of PEMs and enhance the polymer chain mobility; (ii) employ inorganic fillers or other additives to increase free ions in PEMs; (iii) adjust the ion transport pathways to minimize the distance ions travel within the PEMs; and (iv) optimize the ion-conducting group framework of PEMs.^45,46 Indeed, the iterative update of PEMs is also based on these four directions. From another perspective, the advantages of PEMs in high-density energy storage and practical applications are primarily reflected in the following five areas, including the improvement of ionic conductivity, the increase of voltage stability, the inhibition of dendrite formation, the enhancement of mechanical properties, and the enhancement of interfacial compatibility (Fig. 2b). Briefly, the ionic conductivity of PEMs can be enhanced by strengthening ion-pair dissociation (increasing the dielectric constant of the polymer), reducing glass transition temperature (T_g) (crosslinking the polymer chains), and decreasing the crystallinity of the polymer (adding nanofillers).^47–49 The voltage stability of polymer electrolytes can be improved by using compatible polymer layers or in situ formation of an artificial anionic interphase. The structural and mechanical properties of PEMs can also be designed to physically inhibit dendrite growth, for example, achieving the limitation of ion deposition in crosslinked polymer electrolytes and inhibiting dendrite growth in block copolymers. That is, the ideal PEMs are required to possess excellent electrochemical/chemical/thermal stability, outstanding interfacial compatibility, high ionic conductivity, superior mechanical properties, and high ionic transference numbers to better meet the requirements of high-power energy device applications.^50,51

For PEMs in a broad sense, their design should focus on enhancing ionic conductivity, optimizing interfacial design, and achieving specific performance requirements of corresponding electrochemical devices through structural adjustments. Notably, T_g, ion-pair dissociation, and crystallinity remain the most critical parameters to consider in the design of high-performance PEMs. In this regard, strategies such as macroscopic morphology design of PEMs, microscopic structure design of polymer chains, and regulation of host–guest relationship between PEM matrices and additives can be employed to iteratively develop novel PEMs. Indeed, trade-offs are nearly always present when designing a single PEM that incorporates all desired characteristics, making composite materials or multiphase systems appear to be more promising avenues for meeting these requirements.

2.2 Opportunities and potential of cellulose

2.2.1 Sources and types. Cellulose is a ubiquitous fibrous biopolymer with the empirical formula (C₆H₁₀O₅)_n that can be isolated from natural sources such as trees, wheatgrass, cotton, or bamboo.^21,52–54 Structurally, cellulose is a linear homopolysaccharide characterized by its anhydro-D-glucose repeating units. The repeating unit of cellulose typically consists of two anhydroglucose units (AGU, its empirical formula is (C₆H₁₀O₅)_n, where n is the degree of polymerization) covalently linked by a β-1-4 glucosidic bond, wherein the C₁ position of one glucose ring is connected to the C₄ position of the next glucose ring (Fig. 3a).^34,55 The intra-chain hydrogen bonds formed between the hydroxyl group on C₃ and ring O (C₃–H⋯O) of the adjacent AGU and between the hydroxyl group on C₂ and –OH groups on C₆ of adjacent AGU (C₂–H⋯O) collectively constitute the linear structure of the cellulose chain. Instead of existing as isolated macromolecules in nature, cellulose occurs as aggregates of individual cellulose chains. Depending on the diameter, longitudinal dimensions, types, and treatment of the individual cellulose fibrils, these materials are generally classified as raw cellulose, cellulose microfibers (CMFs), cellulose nanofibers (CNFs), cellulose nanocrystals (CNCs), and several cellulose derivatives (Fig. 3b).^56–58 These types of cellulose materials have different morphologies, crystal structures, porosities, and functional groups.

Fig. 3 An overview of cellulose in the advancement of PEMs. (a) Schematic illustration of the hierarchical structure of cellulose fibers; (b) schematic illustration of the molecular engineering strategies for cellulose (reproduced with permission from ref. 34. Copyright 2025, Springer Nature); (c) structural features and general properties of cellulose and their benefits for PEMs.

Raw cellulose refers to the cellulose that is fabricated without chemical modification and nano-processing of raw materials. It is an important biopolymer found in nature, mainly in the cell walls of plants to provide structural support and stability, and its main sources are wood, cotton, flax, hemp, bamboo, and other plants. CMFs are tiny fibrous structures composed of cellulose molecules.⁵⁹ These microfibers are usually derived from cellulose in plant cell walls, and the original cellulose material is broken down into small and delicate fibers by mechanical or chemical treatment. CMFs typically range in diameter from tens to hundreds of microns and in length from a few millimeters to tens of millimeters. CNFs refer to cellulose structures with nanoscale dimensions, where the diameter or thickness ranges from several to tens of nanometers.⁶⁰ CNFs typically have high specific surface area, specific length, mechanical strength, and flexibility. CNCs are nanofibers that are chemically extracted from plant cellulose.^26,61,62 This process usually involves acid hydrolysis or alkaline treatment of cellulose, which removes the amorphous regions while preserving the crystalline regions to obtain nanofiber particles. Therefore, the crystallinity can reach 90%, and the resulting nanofibers possess high mechanical strength and rigidity. In addition, various cellulose derivatives (e.g., methyl cellulose, ethyl cellulose, carboxymethyl cellulose, hydroxyethyl cellulose, cellulose acetate, hydroxypropyl cellulose, cellulose sulfate, and cellulose nitrate) have been developed and utilized by using chemical treatment or functionalization for upgrading the value or expanding versatility.^63–65 Briefly, the properties of cellulose derivatives depend not only on the type and degree of substitution but also on the mode of functionalization on the polymer chain. In addition, several molecular engineering strategies were employed to chemically modify cellulose, such as etherification, oxidation, and esterification (Fig. 3b).⁶³ Briefly, etherification involves replacing the hydrogen atom on the hydroxyl group with an alkyl chain to make the cellulose water-soluble or swellable. Oxidation introduces oxygen-containing functional groups, such as carboxyl groups, which can improve dispersion and solubility. Esterification causes the cellulose hydroxyl group to react with a carboxylic acid or its derivatives to form an ester, thus modifying properties such as reactivity, thermal stability, or hydrophobicity, depending on the specific type of functional group introduced.

2.2.2 Structures and properties. As we mentioned above, cellulose is composed of the basic unit glucose, and the connection between glucose molecules is achieved by the β-1,4-glucoside bond. During this glycosidic bond formation process, the C1 carbon atom of one glucose molecule and the C4 carbon atom of another glucose molecule are catalyzed by an acid to form a connection.⁶⁶ Cellulose chains are usually unbranched, and hydrogen bonds within the chains form stable connections between hydrogen bonds and oxygen of adjacent molecules, resulting in a highly linear chain-like arrangement. Moreover, some of the inherent properties of cellulose are also worth discussing briefly, and these will also be covered in subsequent discussions. For example, cellulose is insoluble in water and common solvents due to the fact that cellulose molecules interact through intramolecular and intermolecular hydrogen bonding as well as electrostatic and hydrophobic interactions within the integrated fibers, which increase the limitations of its applications to some extent. Therefore, the development of suitable solvents to dissolve cellulose has received extensive attention in previous studies. In this regard, ionic liquids (e.g., 1-butyl-3-methylimidazolium chloride, 1-butyl-3-methylimidazolium formate, N,N-dimethylacetamide/lithium chloride, NaOH/thiourea, LiOH/urea, and NaOH–urea) with low toxicity, thermal stability, negligible volatility, and recyclability were employed as effective solvents for cellulose. From another perspective, the superiority of cellulose is mainly reflected in its remarkable inherent stiffness and strength. For instance, CNFs exhibited excellent intrinsic mechanical properties due to high crystallinity and intermolecular interactions. In addition, cellulose also shows some hygroscopic properties due to its tendency to form hydrogen bonds with water. In this process, water molecules were attracted and immobilized in the cellulose structure and water penetration was mainly confined to the disordered structural domains.^67,68

Indeed, abundant sources, good hygroscopicity, excellent mechanical properties, superior thermal stability, and editable structural features have enabled cellulose to show unprecedented potential for the construction of PEMs and have inspired enduring research enthusiasm.^69–71 A deep understanding of the relationship between cellulose properties and functions is essential to fully leverage its potential in PEM design, as the functionality of materials typically depends on their intrinsic structure and corresponding physicochemical characteristics.^72–74 Briefly, the relationship between cellulose structure, properties, and benefits in PEM construction is summarized as follows (Fig. 3c and Table S2): (i) the naturally abundant and biodegradable nature of cellulose makes it an ideal candidate for developing novel PEMs to replace existing petrochemical feedstocks; (ii) the abundant polar hydroxyl groups in cellulose can be used to introduce various functional groups for flexible performance enhancement; and (iii) the abundant hydroxyl groups form strong intermolecular and intramolecular hydrogen bonds, endowing cellulose fibers with excellent mechanical properties.⁷⁵

3. Molecular engineering strategies of cellulose for PEMs

3.1 Fabrication strategies of cellulose-mediated PEMs

During the preparation of cellulose-mediated PEMs, an interconnected network structure composed of cellulose chains is typically constructed. Recently, various fabrication strategies have been presented based on the characteristics of the starting cellulose materials, including solution casting/coating, vacuum-filtration, electrostatic spinning, phase inversion, and papermaking. Following these approaches, PEMs can be flexibly fabricated using water-soluble cellulose derivatives, dissolved cellulose, BC, and CNF, etc.⁷⁶ In this section, we briefly overview the fabrication strategies of cellulose-mediated PEMs to build a bridge between cellulose and PEMs (Fig. 4a).

Fig. 4 An overview of molecular engineering strategies for cellulose in PEMs. (a) Schematic illustration of the fabrication strategies for cellulose-mediated PEMs (solution casting/coating; vacuum-filtration; electrostatic spinning; phase inversion); (b) molecular engineering strategies for the conversion of cellulose to PEMs (esterification reaction; oxidation reaction; sulfonation reaction; phosphorylation reaction; etherification reaction/addition reaction; ionic-crosslinking; molecular-crosslinking; immersion curing; and molecule bonding).

Solution casting is the most commonly used method for preparing PEMs, as this process allows for easy adjustment of membrane thickness and enables large-area production. In this process, PEMs are cured by evaporating the solvent on an inert mold or by employing the blade casting method.⁷⁷ Notably, considering the poor solubility of original cellulose in most solvents, recent studies have typically employed soluble cellulose derivatives to design high-performance PEMs, such as esterified cellulose and methyl cellulose (the detailed fabrication process will be mentioned in the subsequent section). Despite the extensive attention and research on the solution recasting method, several unavoidable problems still deserve special attention: (i) low membrane formation efficiency; (ii) possible recrystallization of the polymer during prolonged drying; (iii) non-uniformity of membrane formation; and (iv) involvement of organic reagents. Vacuum filtration technology uses vacuum conditions to pump cellulose suspension onto the filter mesh to form PEMs. The thickness and porosity of the resulting membrane can be effectively controlled by regulating the volume and concentration of the filtrate. However, this method may cause two issues: (i) the formed cellulose-mediated PEMs may fit tightly on the filter and be difficult to dislodge due to atmospheric pressure and (ii) the two sides of PEMs may have different pore structures. Moreover, electrospinning is a special nanofiber fabrication technique for cellulose-mediated PEMs. Different from other methods, the electrospinning method allows the processing of membranes consisting of micro-/nano-sized fibers.⁷⁸ In this regard, the randomly oriented fibers were interconnected to form a three-dimensional network with a porosity of more than 80%. Notably, the electrostatic spinning technique has some limitations in the selection of cellulose materials. For example, most of the reported cellulose acetate-mediated PEMs have been prepared by electrospinning, and they have excellent mechanical properties, outstanding electrolyte uptake, and high melting temperature. Additionally, electrospinning is performed under a strong electric field, producing fibers through electrostatic traction, with the drawbacks of low spinning rates and high production costs. Therefore, the development of spinning methods without an electric field (e.g., force spinning, melt spinning, and wet spinning) may be a future direction. Phase inversion involves the transformation of a polymer from a solution to a three-dimensional porous solid state, with accompanying changes in solvent and non-solvent. Specifically, phase inversion can occur through multiple methods, including non-solvent induced phase separation, solvent-evaporation phase inversion, vapor-deposition phase inversion, and thermally induced phase separation.⁷⁹ For example, porous CMC-mediated PEMs can be prepared using a solvent-free evaporation technique, and the pore structure of the PEMs can be adjusted by varying the volume ratio of DMF solvent. Furthermore, the papermaking process is also one of the traditional methods for fabricating cellulose-mediated PEMs. In this method, cellulose fibers are suspended in water and formed into a fiber network on a mesh support. In brief, the papermaking process allows for large-scale production of cellulose-mediated PEMs for low-cost battery applications.

3.2 Molecular engineering strategies of cellulose-mediated PEMs

Despite the extensive research interest in PEMs, practical considerations such as environmental friendliness, cost-effectiveness, and processability must be taken into account to ensure that these next-generation PEMs have long-term commercial applicability.^80,81 The abundance of hydroxyl groups on the cellulose skeleton offers a multifunctional platform for molecular design. Based on the implementation of various construction strategies, cellulose has been successfully prepared as high-performance PEMs. Although natural cellulose holds great potential as PEMs, its ionic conductivity is still limited by the high crystallinity of the molecular structure and the high glass transition temperature. In this regard, converting inert natural cellulose into high-performance PEMs is possible through several molecular engineering strategies. In this section, we discuss the molecular engineering strategies of cellulose and the design concepts of cellulose-mediated PEMs in detail and critically combined with specific examples (Fig. 4b). In addition, the properties of cellulose-based PEMs prepared using these molecular engineering strategies are displayed in Table S3.

3.2.1 Ion-conducting groups grafted by chemical engineering.
Etherification/addition reaction. Along different design lines, Zhou and coworkers developed an amphoteric cellulose-mediated PEM (CEQC) using Williamson etherification and Michael addition, which can effectively suppress side reactions and dendrite growth in aqueous zinc batteries, resulting in stable Zn anodes (Fig. 5a).⁸² On the basis that amphiphilic groups can modulate the electrodeposition behavior and change the solvation structure of Zn²⁺, amphiphilic cellulose with positively charged [–(CH₃)₃N⁺] and negatively charged –COO⁻ groups was introduced as a PEM by using the abundance of modifiable hydroxyl groups on the cellulose chain. Specifically, this process involved the reaction of 3-chloro-2-hydroxypropyltrimethylammonium chloride and acrylamide with –OH groups of cellulose (the reactivity order: C3 < C2 < C6) via Williamson etherification and Michael addition. Notably, the number of –(CH₃)₃N⁺ and –COO⁻ groups on the cellulose molecule chain can be effectively controlled by adjusting the reaction conditions (e.g., the feed ratio and contact time). Briefly, the negatively charged –COO– group in CEQC can replace the H₂O molecule in ZnSO₄ and form an ionic complex with Zn²⁺, thus lowering the de-solvation barrier and shortening the transfer path of Zn²⁺. The positively charged –(CH₃)₃N⁺ group in CEQC can be adsorbed on the Zn anode to form a stable and homogeneous protective layer, preventing H₂O from coming into direct contact with the Zn surface, thus promoting the uniform deposition of Zn²⁺. Therefore, the synergistic effect of CEQC resulted in a long cycling life of the Zn//Zn symmetric battery over 5000 h (0.5 mA cm⁻² and 0.5 mA cm⁻²), high DOD_Zn (200 h with a Zn utilization rate of 68.4%), and high CE (∼99.6% over 1500 cycles at 2 mA cm⁻²).

Fig. 5 Strategies of cellulose conversion to PEMs through chemical engineering. (a) Schematic diagram of the preparation process of CEQC from cellulose by etherification and addition reaction and the galvanostatic cycling of Zn plating–stripping in Zn//Zn symmetrical batteries (reproduced with permission from ref. 82. Copyright 2025, American Chemical Society). (b) Schematic of the fabrication process of CCNF prepared by TEMPO oxidation reaction and the cycling curves of Zn plating–stripping in Zn//Zn symmetrical batteries (reproduced with permission from ref. 84. Copyright 2024, Wiley-VCH). (c) Schematic diagram of grafting phosphate groups onto the cellulose backbone and the Li plating/stripping behavior of Li//Li symmetric batteries (reproduced with permission from ref. 86. Copyright 2025, Springer Nature). (d) Schematic diagram of the fabrication process of sulfonated cellulose prepared by sulfonation reaction and the galvanostatic cycling of Zn plating–stripping in Zn//Zn symmetrical batteries (reproduced with permission from ref. 87. Copyright 2024, Wiley-VCH). (e) Schematic illustration of the replacement of the hydroxyl group of native cellulose with the phthalate group by esterification reaction and the cycling stability of Li//CP//Li symmetric batteries (reproduced with permission from ref. 88. Copyright 2024, Springer Nature). (f) Schematic diagram of the fabrication process of CLA-CN from CLA by condensation reaction and the cycling stability of Li//Li symmetric batteries (reproduced with permission from ref. 89. Copyright 2024, Wiley-VCH).

In general, exceptional ionic conduction properties are the primary factor to be considered for cellulose-mediated PEMs. As we mentioned above, PEMs with different functions and properties can be constructed by multidimensional molecule engineering of cellulose, such as esterification reaction, oxidation reaction, sulfonation reaction, phosphorylation reaction, and etherification reaction/addition reaction. In this regard, different methods of grafting ionic groups enabled the prepared cellulose-mediated PEMs to exhibit unique properties, such as physical properties in the macroscopic view and chemical properties in the microscopic view. It can be said that new high-performance PEMs can be obtained by grafting different ionic/molecular groups on the cellulose backbone by chemical grafting. Notably, the successful realization of molecule engineering of cellulose demands a suitable solution system to dissolve cellulose.⁸³ For example, in etherification and addition reactions of cellulose, the cellulose needs to be fully solubilized in the base/urea system to complete the subsequent reaction. In other words, the solubilization strategy of cellulose is a prerequisite for its molecular engineering, which is explained in detail in Section 3.2.4.

Oxidation reaction. Although relatively satisfactory ionic conductivities have been obtained for cellulose as a quasi-solid electrolyte, its low ionic conductivity and mobility remain a challenge when used as an all-solid electrolyte. Therefore, Xu and coworkers prepared a carboxylate cellulose-mediated PEM (Zn-CCNF@XG) with a wide electrochemical window, high ionic conductivity (1.17 × 10⁻⁴ S cm⁻¹ at room temperature), and low cost for a zinc-ion battery by oxidation reaction of cellulose fibers (Fig. 5b).⁸⁴ The TEMPO-mediated oxidation method selectively oxidized the –OH group at the C₆ position to a –COOH group while keeping the integrity of the cellulose structure. Notably, the molecular electrostatic potential of CCNF indicated that the negative charge centers were concentrated on the oxygen-containing functional groups (–O–, –OH, and –COOH) serving as Zn²⁺ transition sites. That is, –OH and –COOH synergistically constructed the nanochannel for fast migration of Zn²⁺, which led to the superior Zn²⁺ transport performance of Zn-CCNF@XG. Furthermore, the negative charge centers were mainly concentrated on oxygen-containing functional groups, suggesting that these groups function as Lewis basic sites and may provide ionic jumping sites for the migration of Zn²⁺ in cellulose-mediated PEMs. Moreover, the Zn-CCNF@XG PEM was constructed by co-mixing CCNF and XG and then heating at 90 °C by solution-casting. In particular, the Zn//Cu symmetric battery based on Zn-CCNF@XG exhibited excellent coulombic efficiency retention (99.51%) after 1000 cycles, and low overpotential (∼40 mV) was observed after 1800 cycles (3600 h) for a Zn//Zn symmetric battery at 0.5 mA cm⁻² and 0.5 mAh cm⁻². Notably, different from the disordered motion in liquid electrolytes, Zn²⁺ can migrate uniformly along the cellulose molecular chains, which facilitates the uniform deposition of Zn²⁺. Meanwhile, the ultra-high mechanical stress and dense structure of Zn-CCNF@XG can also effectively prevent the formation of Zn dendrites.

Following a similar research approach, Wen and coworkers developed a sulfonation strategy to prepare functionalized PEMs for LiI-involved Li–O₂ batteries by grafting negatively charged –SO₃⁻ groups on bacterial cellulose (BC) to gel with liquid electrolytes.⁸⁵ Thanks to the high density distribution of negatively charged –SO₃⁻ on sulfonated BC, the functionalized PEMs exert strong electrostatic repulsion on the negatively charged I₃⁻, thereby suppressing the shuttle effect. In other words, the electrostatic repulsion between negatively charged sulfonated BC PEMs (SBC) and I₃ effectively inhibits the redox shuttle reaction. Notably, the SBC was fabricated via a two-step method. In brief, BC underwent oxidation in the presence of an oxidizing agent, wherein the C2,3-hydroxyl group in the repeating unit was oxidized to an aldehyde group to produce the 2,3-dialdehyde BC (DBC). The DBC was then processed with NaHSO₃, during which sulfite ions were attached to the aldehyde group and the sulfonic acid groups were grafted to obtain SBC. The SBC was finally immersed in a DMSO-based electrolyte to achieve gelation of the polymer backbone and generate SBC PEMs (SBCG). Moreover, the SBCG exhibited high ionic conductivity at room temperature (1.6 × 10⁻³ S cm⁻¹), comparable to liquid electrolyte/GF (2.58 × 10⁻³ S cm⁻¹). The high ionic conductivity of SBCG can be attributed to the complete gelation of the 3D network structure and the full Li⁺ solubilization resulting from the interaction with the negatively charged –SO₃⁻. Furthermore, the Li–O₂ battery with SBCG not only inhibited the I₃⁻ shuttle activity, but also had better cycling performance within 60 cycles.

Phosphorization reaction. Previous studies have primarily targeted the regulation of electrostatic interactions, dipole–dipole interactions, and van der Waals forces between the PEMs and electrolyte components (e.g., anions, alkali metal ions, and solvent molecules). Nevertheless, their electrical properties have long been overlooked, especially in the presence of the strong electric fields within electrochemical equipment. Therefore, to investigate the response behavior of cellulose-mediated PEMs in electric fields at different dielectric constants, Zhu and coworkers introduced a dielectric constant as a descriptive parameter and developed an ultra-high dielectric constant fiber PEM constructed primarily from phosphoric acid-treated cellulose (PC) (Fig. 5c).⁸⁶ Generally, the PEMs with high dielectric constants are crucial for regulating the electric field distribution and preventing tip effects at the metal anode, thereby extending the battery life. Thus, polar phosphate groups were selected for introduction into cellulose to enhance its dielectric constant. Specifically, the phosphate groups were grafted onto the cellulose backbone by treating the cellulose in an (NH₄)₂PO₄ solution. The addition of urea to the reaction system can inhibit the release of protons from (NH₄)₂PO₄, thus avoiding cellulose degradation. Remarkably, PCS with a higher dielectric constant and ionic conduction property (ionic conductivity was 0.76 mS cm⁻¹ at 25 °C and Li⁺ transference number is 0.68) exhibited significantly greater electron transfer and dipole moment increments compared to CS, indicating stronger interactions between PCS and Li⁺ under an electric field. Therefore, the Li//Li symmetric batteries equipped with PCS exhibited no significant cell short circuiting or voltage polarization (overpotential is 51.3 mV) after prolonged cycling for over 1600 h at 1 mA cm⁻² and 1 mAh cm⁻². Moreover, the Li//LiFePO₄ pouch cell with PCS achieved high specific energy (350 Wh kg⁻¹) at practical quantities of active materials and electrolyte.
Sulfonation reaction. In general, separators containing anionic functional groups show great potential for mitigating parasitic reactions involving electrolyte anions and facilitating parallel Zn deposition. Apart from the chemical modification treatment of cellulose mentioned above, the sulfonation reaction also serves as an effective strategy for improving the ion transport kinetics of cellulose-mediated PEMs. In this regard, Zhang and coworkers designed a sulfonated cellulose PEM (CF-SO₃) with high mechanical strength, low thickness, and high ionic conductivity for aqueous zinc-ion batteries (Fig. 5d).⁸⁷ Notably, the chemically grafted –SO₃⁻ groups on the cellulose backbone can effectively inhibit the migration of SO₄²⁻ anions due to the existence of electrostatic repulsion, thus increasing the ion transfer number of Zn²⁺. Moreover, the –SO₃⁻ groups can strongly interact with Zn²⁺, thereby facilitating the desolvation process of hydrated Zn²⁺ and limiting the planar diffusion of Zn²⁺ at the surface of the Zn electrode. The free-standing CF-SO₃ separator with a thickness of about 50 µm can be obtained by vacuum filtration, which is lower than the thickness of the commercial glass fiber (GF) separator (204 µm). In particular, the CF-SO₃ separator also exhibited excellent tensile strength (18.7 MPa) and ionic conductivity (52.1 mS cm⁻¹). With these benefits, the Zn//Zn symmetric cells with the CF-SO₃-mediated PEM can achieve an impressive cycling lifespan of 1200 h at current densities of 2 mA cm⁻² and 4 mAh cm⁻², whereas the CF separator suffered a short circuit at around 384 h (400 h lifespan at a large depth of discharge of 68.3%). In addition, the CF-SO₃ separator also exhibited an excellent rate capability of 90.8% (a discharge capacity of 218.2 mAh g⁻¹ after 4000 cycles) and cycling stability in both highly loaded Zn//MnO₂ and Zn//HAVO full cells.
Esterification reaction. To convert sustainable cellulose into a robust PEM, Cao and coworkers employed a homogeneous esterification reaction to replace the hydroxyl group of cellulose with phthalic anhydride, thereby transforming cellulose into cellulose phthalate (CP) with excellent mechanical (12 MPa) and ionic conductivity (1.09 × 10⁻³ S cm⁻¹) (Fig. 5e).⁸⁸ In this process, the introduction of phthalate groups can facilitate the multi-coordination of Li⁺, which is attributed to the presence of neighboring carboxyl and carbonyl groups on the benzene ring. Notably, the parallel molecular chain conformation in native cellulose fibers has been totally altered in CP-based PEMs, exhibiting a uniformly dispersed and randomly entangled non-oriented polymer network structure. In other words, this conformation enhanced the coupling/decoupling of Li⁺ and formed an efficient transport channel, creating a robust Li⁺ conduction network within the CP. Moreover, the introduced carboxyl groups can form strong hydrogen bonding networks between CP chains, which can enhance the mechanical strength and durability even under conditions of high lithium salt content. Notably, the CP-based PEM was prepared by mixing the obtained CP with a lithium salt and using ethyl carbonate as a plasticizer by solution-casting. In addition, the CP-based PEM also has excellent tensile strength (12 MPa) and thermal stability (decomposition temperature is 274 °C) to better meet the requirements of the operating environment of a Li-ion battery. The Li plating/stripping behavior of the CP-based PEM was further tested by galvanostatic cycling at 0.1 mA cm⁻². The Li//CP//Li symmetric cell exhibited a polarization voltage of 25 mV, indicating a robust and stable interface capable of uniform Li deposition/stripping while inhibiting Li dendrite formation. Moreover, the CP-based PEM has also shown exceptional properties in full cell performance evaluations. For example, an initial CE value of 98.0% and a reversible capacity of up to 158.5 mAh g⁻¹ were achieved during charging at a 0.1C rate, demonstrating performance comparable to that of liquid electrolyte-based batteries (158.4 mAh g⁻¹). Moreover, the Li//CP//LiFePO₄ batteries assembled based on CP achieved a reversible capacity of 157.1 mAh g⁻¹ at a commercially active material loading of 18.3 mg cm⁻², demonstrating excellent potential for stable long-term cycling.
Condensation reaction. Apart from the examples discussed above, cellulose-mediated PEMs can also be effectively constructed using grafting chain extension strategies for cellulose structures. The high temperatures during the heat treatment of batteries can lead to electrolyte decomposition and electrode material damage, which can reduce battery performance and life. Therefore, Tian and coworkers presented an ion-conducting molecular grafting strategy that enhanced Li⁺ transport (1.25 × 10⁻³ S cm⁻¹ at room temperature) by extending cellulose molecules and stabilized the interface for ultra-long cycling (Fig. 5f).⁸⁹ The isocyanate and the hydroxyl group of cellulose acetate (CLA) can react in one step to form a small molecule structure (e.g., –COOH, –CF₃, and –CN) that creates a urea bond to enhance Li⁺ conduction. By comparing the binding energies between different CLA groups and Li⁺, the results showed that the grafted cyano CLA has the lowest binding energy with Li⁺. The lower binding energy allowed the electrolyte to effectively adsorb and deintercalate Li⁺, thus improving the ion transport performance of lithium-ion batteries. More specifically, the presence of cyano could enhance the interaction between the polymer chains and ions, helping the ions to move through the electrolyte. Moreover, small molecules with different structures were grafted to replace the hydroxyl group on CLA and then mixed with Li_1.3Al_0.3Ti_1.7(PO₄)₃ (LATP) ceramic particles to prepare CLA-x-LATP PEMs with different branched groups by the thermoforming method. Benefiting from the grafted cyano-functional molecules, the Li//CLA-CN-LATP//Li symmetric battery showed low polarization voltage and high cycling stability, with a Li stripping/plating cycle of more than 1200 h at 0.1 mA cm⁻². Meanwhile, the Li//CLA-CN-LATP//Li symmetric battery can maintain stable and low polarization at the current density of 0.54 mA cm⁻², achieving a critical current density value of 0.64 mA cm⁻². In addition, CLA-CN-LATP in the LiFePO₄//CLA-CN-LATP//Li battery also had a cycle stability of up to 1500 cycles with 92.1% capacity retention at 0.5C and 25 °C.

3.2.2 Ion-conducting channels constructed by cross-linking.
Ionic crosslinking. As previously noted, cellulose is abundantly available from diverse biomass sources and exhibits a uniaxially aligned, one-dimensional hierarchical structure, characterized by a high density of oxygen-containing polar functional groups, originating from the repeating anhydroglucose units (AGUs) that constitute the molecular chains of cellulose.⁹⁰ Notably, the polar functional groups of cellulose can solvate lithium ions and acid during lithium ion migration. However, the narrow spacing between cellulose molecular chains adversely affects lithium ion doping. In other words, cellulose was previously used as an inert supporting matrix for gel/ liquid electrolytes or other PEMs.^91–93 Therefore, to solve this dilemma, Hu and coworkers fabricated a high-performance cellulose nanofiber (CNF)-based solid polymer ionic conductor (Li–Cu–CNF) by coordination of Cu²⁺ with one-dimensional cellulose nanofibers (Fig. 6a).⁹⁴ The Li–Cu–CNF ionic conductors were fabricated through simple ion coordination and solvent exchange processes. In this process, the coordination interaction between Cu²⁺ and cellulose can form molecular channels by extending the interpolymer chain spacing, thereby altering the crystalline structure of cellulose and enabling lithium ions insertion and rapid transport. The pristine cellulose exhibits a typical I_β-type cellulose monoclinic diffraction pattern with a closely linked cellulose molecular distance of 0.39 nm. In contrast, Cu–CNF–NaOH treated with Cu²⁺-saturated alkali solution had a hexagonal crystal structure with triple symmetry along the cellulose chain direction. In the simulated Cu–CNF–NaOH structure, each cellulose chain connects to three adjacent cellulose chains with an interchain distance of 0.87 nm, bridged by Cu²⁺ through four-coordinate Cu–O bonds, thereby disrupting the compact molecular packing of cellulose. Benefiting from the cross-linking of Cu²⁺ with cellulose molecular chains, the prepared Li–Cu–CNF PEM exhibited excellent Li⁺ conductivity (1.5 × 10⁻³ S cm⁻¹ at 25 °C), Li⁺ transference number (0.78), and electrochemical stability (0–4.5 volts). In particular, the Li–Cu–CNF PEM can maintain stable Li cycling performance in a Li//Li symmetric battery for 300 h at a current of 0.5 mA cm⁻² without dendrite-induced short circuits. Moreover, long-term cycling of Li–Cu–CNF electrolyte was also achieved in solid-state batteries constructed with Li metal anodes and NMC811 or LiMn₂O₄ cathodes.

Fig. 6 Strategies of cellulose conversion to PEMs through cross-linking and molecule blending. (a) Schematic representation of the coordination of Cu²⁺ with cellulose nanofibers for the preparation of ionic conductors and cycling stability of Li–Cu–CNF electrolyte (reproduced with permission from ref. 94. Copyright 2021, Springer Nature). (b) Schematic representation of the interactions between the components during the preparation of the PLCZ PEM and cycling stability of Zn//Zn symmetric batteries (reproduced with permission from ref. 96. Copyright 2025, Springer Nature). (c) Schematic diagram of the ion transport mechanism and cycle stability of CP@PPC (reproduced with permission from ref. 97. Copyright 2023, Wiley-VCH). (d) Schematic diagram of a CMC-cellulose composite membrane for enhancing the cycling performance of Zn//Zn symmetric batteries (reproduced with permission from ref. 98. Copyright 2023, Wiley-VCH).

Molecular crosslinking. In addition to ion crosslinking, molecular crosslinking is another effective approach for constructing cellulose-mediated PEMs.⁹⁵ To advance the application of cellulose materials in aqueous zinc ion batteries, along a different design concept, Huang and coworkers designed a quasi-solid polymer electrolyte (PLCZ) with superior mechanical properties, fast Zn²⁺ diffusion, and high selectivity by crosslinking cellulose nanofiber (CNF), lithium magnesium silicate (LMS), and sodium polyacrylate (PAAS) (Fig. 6b).⁹⁶ Briefly, LMS and PAAS were dissolved in CNF solution to create a physical entanglement network and after drying it was immersed in ZnSO₄ solution for complete Zn²⁺ exchange to obtain PLCZ PEMs. In this process, the –COO⁻ groups in PAAS can both establish physical connections with CNF and polymerize with LMS sheets by electrostatic interaction and the metal coordination bonds between PAAS and Zn²⁺. Thanks to the complex interaction forces within the system, PLCZ PEMs exhibited excellent structural integrity with no visible damage even after kneading, emphasizing their superior restorability and flexibility. The PLCZ PEMs featured a consistent and sleek surface and a dense cross-sectional texture, which contributed to the interfacial stability between the electrolyte and the Zn anode and created an effective bridge for the rapid transport of Zn²⁺. PLCZ-based Zn symmetric batteries showed an exceptionally long lifetime of more than 4400 h at 0.5 mA cm⁻² and 0.25 mAh cm⁻², nearly 10 and 22 times that of PCZ (without LMS) and CZ (CNF), respectively. Furthermore, the PLCZ PEM also exhibited a cycle life of up to 1400 h without significant voltage fluctuations at 10 mA cm⁻² and 1 mAh cm⁻². Moreover, the Zn//Cu symmetric batteries using the PLCZ PEM achieved a cycle life exceeding 2400 cycles, with a stable average coulombic efficiency as high as 99.7%.

3.2.3 Ion conduction channels optimized by molecular entanglement.
Immersion curing. The abundant hydroxyl functional groups in natural cellulose fibers produce strong intermolecular forces, leading to a reduction in the pore size of the separator and thereby hindering ion transport through the pores. Therefore, grafting chemical functional groups on cellulose fibers and employing ionic/molecular cross-linking are effective ways to construct cellulose-mediated PEMs. Remarkably, grafting more reactive hydroxyl sites onto the cellulose surface can effectively mitigate the side effects caused by hydroxy groups. However, the randomness and complexity of the cross-linking and grafting process require more stringent production and handling processes, making them unsuitable for the industrial production of cellulose-mediated PEMs. In view of the above discussion, Wu and coworkers prepared a cellulose-based PEM (CP@PPC) by impregnating and curing a cellulose-based separator (CP) in poly(propylene carbonate) (PPC) for sodium-ion batteries (Fig. 6c).⁹⁷ Notably, two different ion transfer modes exist in the amorphous and crystalline regions of the PPC chain. Specifically, the PPC chains can form a helical conformation in the crystal region, and Na⁺ is transported directionally by interacting with carbonyl groups within the helical space in the presence of an electric field. The carbonyl group on the PPC chain segment first coordinated with Na⁺ in the amorphous region, and Na⁺ was then continuously “uncoordinated–coordinated” between different chain segments under the movement of the PPC chain segment and the electric field, which led to the continuous transport inside the PPC. In particular, the “pore-hopping” ion transport mechanism in CP@PPC creates additional Na⁺ migration pathways, thereby achieving a high Na⁺ transfer number (0.613) and high ion conductivity (0.686 mS cm⁻¹ at 25 °C). Notably, the full cells with CP@PPC can maintain a superior capacity retention at 2C after 500 cycles compared to CP. Therefore, the initial discharge capacity and cycling stability of the NVPOF//CP@PPC//CC full cell are higher than that of the NVPOF//CP//CC full cell, indicating that CP@PPC holds broad application prospects in the field of sodium ion batteries.
Molecule bonding. Following a similar research line, Hu and coworkers fabricated a nanocellulose–carboxymethylcellulose (CNF/CMC) composite PEM with low free water content, high mechanical properties, and high ionic conductivity for long-cycle-life aqueous zinc ion batteries (Fig. 6d).⁹⁸ Compared to cellulose, CMC with a large number of carboxyl groups can create stronger bonding interactions with water molecules, which contributes to limiting the free water in the PEMs. In other words, the increased bound water content enhances the conductivity of Zn²⁺ while preventing parasitic side reactions with excess free water molecules. Notably, the tensile strength of the CNF/CMC-based PEM could be increased to 70 MPa using a simple NaOH treatment. Therefore, the CNF/CMC-based PEM showed a Zn²⁺ conductivity of up to 26 mS cm⁻¹ and was capable of stable plating/stripping on a Zn anode at 80 mA cm⁻². Moreover, side reactions triggered by free water were also greatly reduced, such as hydrogen evolution and the formation of passivating Zn₄SO₄(OH)₆·xH₂O. In particular, a symmetric battery with CNF/CMC-based PEM exhibited a polarization voltage of 50 mV at 10 mA cm⁻² and 5 mAh cm⁻², remaining constant over 350 cycles. In contrast, an aqueous zinc ion battery with glass fiber separators had an initial polarization of up to 100 mV and a sharp increase in polarization after 30 cycles. Moreover, Zn//Zn symmetric batteries assembled with CNF/CMC also demonstrated stable cycling at 80 mA cm⁻² and 40 mA cm⁻², achieving cumulative plating capacities of 14 and 17.6 Ah cm⁻², respectively. Therefore, cellulose–CMC electrolyte with good electrochemical performance, low cost, ease of preparation, and sustainability holds broad application prospects in zinc-ion batteries.

By employing multi-dimensional, multi-scale molecular engineering strategies, the application of cellulose in polymer electrolyte materials can be effectively achieved. This process requires particular attention to the stability of cellulose molecular chains during molecular engineering treatment, especially the effects of certain acidic oxidizing agents and ionic crosslinking on the structure of cellulose molecular chains.

3.2.4 Physicochemical properties modulated by dissolution–regeneration.
Ionic liquid-mediated dissolution–regeneration. The stable structure and the large number of tightly packed internal hydrogen bonds make cellulose difficult to dissolve in water and common solvents. To some extent, the insoluble nature of cellulose hinders its further design and applications. In this regard, the preparation of gels from cellulose by dissolution regeneration generally involves the steps of dissolution, gelation, solvent replacement and drying. In general, the ability of ionic liquids to dissolve cellulose is because the combined action of anions and cations can disrupt the intermolecular and intramolecular hydrogen bonding structure of cellulose. That is, the stronger the hydrogen bonding between the anion and cellulose in the ionic liquid, the higher the solubility of cellulose within the ionic liquid.^99,100 Therefore, through this ionic liquid-mediated regeneration of cellulose by dissolution, Wang and coworkers reported a cellulose-mediated solvent-residual heterogeneous PEM for high-energy and stable wearable zinc batteries (Fig. 7a).¹⁰¹ Notably, the dual-network cellulose-mediated gel PEMs (DCG) prepared by the dissolution-regeneration method have fast Zn²⁺ transport channels due to polar coordination. Trace amounts of deep eutectic solvents (DES, a mixture of 1-allyl-2,3,4,6,7,8-hexahydropyrrolo[1,2-α]pyrimidin-1-ium chloride and 3-amino-1-propanol) will remain in DCG as functional additives. In particular, the DES in DCG can migrate to the zinc anode under the effect of an electric field, thus displacing the water in the local environment and suppressing side reactions. In contrast, in the pure cellulose gel electrolyte system, Zn²⁺ will accumulate and continue to deposit at the most favorable nucleation sites on the Zn anode surface, thereby generating a tip effect and forming dendrites. In addition, the cycling stability of the DCG-containing Zn anode was investigated using a Zn//Zn symmetric battery. The symmetric batteries with DCG can maintain stable polarization voltage at 4 mA cm⁻² and 4 mAh cm⁻² with a long-term cycle life exceeding 4000 h. The Zn//Zn symmetric batteries employing liquid electrolyte and cellulose-mediated gel electrolyte suffered short circuits after operating for 200 h and 432 h, respectively. In this process, the presence of DES can passivate the zinc planes and direct the deposition of zinc metal along the planes, thus preventing dendrite growth and improving reversibility. Briefly, cellulose can be directly used to prepare high-performance PEMs through an ionic liquid-mediated dissolution–regeneration approach.

Fig. 7 Schematic diagram of regulating the physicochemical properties of PEMs through cellulose dissolution and regeneration. (a) Schematic diagram of the preparation process of DCG and the performance of the Zn//Zn battery (reproduced with permission from ref. 101. Copyright 2025, The Royal Society of Chemistry). (b) Schematic diagram of the preparation process of regenerated cellulose gel electrolyte and the cycling performance of the Zn//Zn symmetric battery (reproduced with permission from ref. 102. Copyright 2025, Elsevier). (c) LiBr/formic acid-mediated dissolution regeneration of cellulose (reproduced with permission from ref. 103. Copyright 2025, American Chemical Society). (d) AlCl₃/ZnCl₂-mediated dissolution regeneration of cellulose (reproduced with permission from ref. 104. Copyright 2022, Royal Society of Chemistry). (e) NaOH-mediated dissolution regeneration of cellulose. (reproduced with permission from ref. 105. Copyright 2024, American Chemical Society).

Alkali/urea-mediated dissolution–regeneration. The alkali/urea-mediated composite system is an effective method for cellulose dissolution and regeneration. Generally, the mechanism of cellulose dissolution can be explained as follows. NaOH exists in aqueous solution as two ionic hydrates, [OH(H₂O)_n]⁻ and [Na(H₂O)_m]⁺. At room temperature, the rapid exchange between free and bound water makes it difficult for [OH(H₂O)_n]⁻ and [Na(H₂O)_m]⁺ to form and maintain new complex structures. At low temperatures, the exchange of free water with bound water slows down so that the structure of the ionic hydrate can be maintained, making it easier for [Na(H₂O)_m]⁺ to combine with hydroxyl groups on cellulose to form a new hydrogen bonding network, which breaks the intramolecular and intermolecular hydrogen bonding of cellulose. The urea or thiourea is not reactive with cellulose molecules, but rapidly encapsulates the outside of the hydrogen bonding network of NaOH and cellulose through dynamic self-assembly, forming a pipeline inclusion that prevents cellulose molecules from self-polymerizing, and it is this form of stable and rapid self-assembly that allows the cellulose to be dissolved rapidly to form a transparent cellulose solution. Therefore, Long and coworkers fabricated a sustainable and recyclable cellulose PEM for aqueous zinc-ion batteries using an alkali/urea-mediated cellulose dissolution regeneration strategy (Fig. 7b).¹⁰² In contrast to commonly used GF separators, the cellulose-mediated electrolyte regenerated by dissolution of the alkali/urea system effectively mitigates water-induced undesirable reactions via hydrogen bonding between free water molecules and abundant –OH groups, enabling highly reversible zinc plating/stripping processes and stabilization of the zinc anode. In addition, the morphology of the zinc anode after cycling showed that numerous zinc dendrites were formed on the zinc anode circulated with aqueous electrolyte in the GF separator and there were many GF inlays, which suggests that the zinc dendrites pierced the GF separator. In contrast, the cellulose-mediated PEM can form a uniform Zn deposition layer on the Zn anode without dendrite formation after cycling. Compared to the GF separator, the Zn//Zn symmetric battery equipped with the cellulose-mediated PEM shows excellent cycling stability over 680 h at 2 mA cm⁻² and 1 mAh cm⁻².

In addition to the above-discussed strategies for modulating the physicochemical properties of PEMs using different cellulose dissolution–regeneration processes, there are other pathways for cellulose dissolution–regeneration, such as LiBr/formic acid-mediated dissolution–regeneration, AlCl₃/ZnCl₂-mediated dissolution–regeneration, and NaOH-mediated dissolution–regeneration. Although these cellulose dissolution–regeneration strategies have not been applied to the construction of cellulose-mediated PEMs, there remains great potential for future development. Although several cellulose solubilization systems have been developed, the majority of them suffer from disadvantages such as process complexity, high cost, high toxicity, and difficulty in recovery. Therefore, in this section, we focus on several cellulose solvent systems that involve mild conditions and are highly efficient, environmentally safe, and easy to recover.

LiBr/formic acid-mediated dissolution–regeneration. Compared to other acids, commercial formic acid is the simple organic carboxylic acid that can be extracted from biomass and is also far less acidic than most inorganic acids, making it less corrosive to equipment. Moreover, formic acid can be efficiently recovered by vacuum distillation due to its lower boiling point to guarantee a clean process. Therefore, Li and coworkers proposed a cellulose dissolution method that combined LiBr·3H₂O pre-impregnation at room temperature with cellulose dissolution in formic acid to enable controlled production of regenerated nanocellulose (RNCF) (Fig. 7c).¹⁰³ In particular, this method can obtain cellulose with high solubility, thus contributing to the fabrication of tunable regenerated formic acid nanocellulose for various applications. Specifically, the LiBr·3H₂O pre-impregnation at room temperature can significantly improve the dissolution efficiency of cotton pulp by increasing the accessibility of cellulose through the disruption of hydrogen bonding and decomposition of cellulose crystalline structure. This effect is ascribed to the coordination interaction between hydrated Li⁺ and hydrogen bonds in cellulose, as well as the synergistic action between Br⁻ and H-bonds in cellulose hydroxyl groups. The subsequent regeneration processes can produce RNCF with a yield of 100%, featuring a high degree of substitution (DS > 1.2) and an adjustable polymerization degree (300–700). Furthermore, the obtained filaments and membranes based on RNCF have a tenacity of 1.4 cN/dtex and a tensile strength of 118 MPa, which was higher than the tenacity of cotton fiber and commercial plastics.
AlCl₃/ZnCl₂-mediated dissolution–regeneration. To overcome the drawbacks of cellulose degradation and decreased polymerization produced by conventional inorganic salt systems when dissolving cellulose, Wang and coworkers reported a binary inorganic salt solvent (AlCl₃/ZnCl₂·4H₂O) capable of dissolving cellulose at high concentrations in short periods of time and enabling green, efficient, and environmentally safe production (Fig. 7d).¹⁰⁴ In particular, the binary inorganic salt solution maintained long-term stability, and the concentration of the regenerated cellulose obtained showed no significant decrease. In this process, the cellulose with a polymerization degree of up to 4080 was rapidly dissolved within a reaction time of 120 min. Notably, only slight degradation of cellulose occurred during the dissolution process (30% reduction in the degree of polymerization), much lower than that for cellulose dissolved at 80 °C in a conventional 65% ZnCl₂ solvent system (70% reduction in the degree of polymerization). Therefore, solutions of binary inorganic salts consisting of metal ions with sufficiently small ionic radii (e.g., AlCl₃/ZnCl₂, MnCl₂/ZnCl₂, FeCl₂/ZnCl₂, and LiCl/ZnCl₂) can both dissolve cellulose at room temperature. Importantly, the recovery of inorganic salts in this binary system can reach 95% and can maintain 64% after ten times of recycling, and the dissolution efficiency is the same as that of fresh solvent, which realizes the greening of the cellulose dissolution process.
NaOH-mediated dissolution–regeneration. In addition to the inorganic salt-mediated cellulose dissolution–regeneration mentioned above, alkaline water is attractive because it is fairly simple and the reagents are easily recovered, inexpensive, and widely used in the pulp industry. Briefly, the hydroxide can break hydrogen bonds in cellulose chains at low temperatures when dissolving cellulose, while metal ions such as Li⁺ and Na⁺ can form hydrates with hydroxyl groups to prevent recrystallization of cellulose. Therefore, along this research direction, Zhang and coworkers developed a method for preparing regenerated cellulose using an aqueous alkali solution to efficiently dissolve cellulose in a few minutes at temperatures above 0 °C (Fig. 7e).¹⁰⁵ Specifically, the raw cellulose was first modified using carboxymethyl groups, and then the modified cellulose was completely dissolved with an aqueous NaOH solution at a temperature above 0 °C. The charged groups can weaken inter- and intramolecular hydrogen bonding in cellulose molecules, thus facilitating cellulose solubilization. The cellulose solution was regenerated in a mixture of ethanol and water to produce a regenerated cellulose material. Notably, the liquid waste generated in this process was distilled to recycle ethanol and the residual aqueous NaOH was thermally condensed to an appropriate concentration for reuse. Compared with the previous NaOH–water methods, this approach has the following advantages: (i) no reagent needs to be added to aqueous NaOH; (ii) moderate dissolution temperature (>0 °C); and (iii) relatively high maximum dissolved concentration (∼14%). Moreover, this approach can be extended to various chemical modifications of cellulose as well as other raw materials containing cellulose.

Although various molecular engineering strategies have been developed to design different cellulose-mediated PEMs, it should be explicitly noted that numerous unresolved issues remain: (i) accessibility and uniformity of reaction sites along the cellulose molecular chain: the crystalline regions of cellulose and its robust intramolecular/intermolecular hydrogen bond network hinder chemical reagents from penetrating and uniformly contacting all hydroxyl reaction sites, causing reactions to predominantly occur in amorphous zones and on crystal surfaces, resulting in uneven graft chain distribution and the formation of an “island structure”. Moreover, cellulose backbone chains are prone to hydrolysis or oxidative cleavage under strong oxidation or sulfonation conditions, leading to reduced molecular weight and compromising mechanical strength. (ii) Long-term electrochemical and physicochemical stability: although cellulose exhibits good thermal stability, its grafted functional groups and trace moisture or impurities remaining in the electrolyte may be affected by the strong reducing properties of alkali metals and high voltage. The strength of ionic crosslinks is influenced by environmental factors, potentially leading to relaxation or destruction of the network structure during long-term cycling or at elevated temperatures, resulting in dimensional instability and performance degradation.

4. Key features and the corresponding construction strategies of cellulose-mediated PEMs

As we mentioned above, despite the great potential of cellulose as a functional biopolymer for a wide range of applications, its multiple properties, including ionic conductivity, porosity, mechanical, and safety are still at a suboptimal level for PEMs. In particular, the cellulose-mediated PEMs have been significantly optimized and enhanced in terms of the core indicators of PEMs (e.g., ionic conductivity, ion-selectivity, porosity, mechanical property, thermal property, dendrite inhibition, flame retardant, self-healing, and photoelectric stability). Therefore, in this section, we provide a comprehensive and detailed discussion of the key features of cellulose-mediated PEMs and the corresponding construction strategies for the core components of PEMs. It is worth noting that the strategies for constructing these key properties are not limited to molecular engineering approaches for cellulose but also encompass aspects involving multicomponent composites, physical processing, and action mechanisms. This section aims to provide a comprehensive overview of the key characteristics of cellulose-mediated PEMs and present detailed construction strategies for these key properties. Moreover, the properties of cellulose-mediated PEMs reported in the literature are summarized in Table S4.

4.1 Ion-transport improvement

The ionic conductivity represents the total sum of mobile charge carriers passing through the electrolyte and can be described by eqn (1).where C_i and μ_i are the concentration and mobility of charge carriers and q_i is the ionic charge. The ionic conductivity of PEMs largely determines the electrochemical performance of electrochemical devices. In general, PEMs are usually employed at elevated temperatures to improve the conductivity of ions, the segmentation of the polymer, and accelerate ionic mobility.^11,106 However, elevated temperatures can reduce the mechanical strength of PEMs and further affect the safety of the electrochemical devices. Other effects to improve the ionic conductivity and ionic transfer numbers of PEMs include modifying the polymer matrix structure with crosslinked polymers, blocking ion-conducting polymers, or incorporating inorganic fillers.⁸⁰ In the case of native cellulose, its high crystallinity can hinder ion transport to some extent as well as ion–polymer interactions during metal salt dissociation. Meanwhile, the high glass transition temperature of native cellulose can limit ionic conduction at room temperature. Therefore, treatment of native cellulose is usually required to achieve prominent ionic conductivity. Notably, in PEO or PVDF-based systems, ions are typically solvated by the polymer chains (e.g., Li⁺ by ethylene oxide units in PEO). The ion transport is primarily coupled to the segmental motion of the polymer chains. The ion conduction process follows the Vogel–Tammann–Fulcher (VTF) behavior, strongly dependent on temperature and polymer chain mobility. Briefly, the amorphous regions of PEO chains undergo thermal motion to form dynamic ion-conducting pathways, which are highly dependent on temperature and plasticization. The semi-crystalline structure of PVDF limits chain mobility, so PVDF typically requires the following modification methods to enable its applications in PEMs, including organic polymer modification (e.g., copolymerization, grafting, and blending), inorganic filler modification (e.g., inert fillers and active fillers), liquid additive modification (e.g., liquid organics and ionic liquids), and structural design (e.g., random distribution, multilayer structure, and 3D skeleton structure).¹⁰⁷ For cellulose-mediated PEMs, the functional groups (e.g., –SO₃H, –COOH, and –PO₃H₂) are covalently attached to the rigid cellulose backbone. The resulting well-defined nanochannels create a percolated pathway for ions. The dominant mechanism is thus fixed-channel hopping (or vehicle mechanism for protons), where ions migrate by hopping between adjacent fixed sites, with the assistance of solvent molecules (especially in hydrated membranes). Notably, the fixed-channel hopping mechanism employed in cellulose-mediated PEMs offers significant advantages, decoupling ion transport from polymer chain dynamics, and holds promise for achieving more stable and predictable performance across a wide temperature range. Therefore, in this section, we focus on strategies to enhance the ion transport of cellulose-mediated PEMs, extending beyond molecular engineering approaches to include multi-component composite strategies.

Grafting of ion-conducting molecules. The strong hydrogen bonding interactions between cellulose molecules cause the molecular chains to pack tightly, thus hindering ionic conduction.¹⁰⁸ To solve this problem, Tian and coworkers proposed an ion-conducting molecular grafting strategy for the preparation of cellulose acetate-mediated PEMs (CLA-CN-LATP) with high ionic conductivity at room temperature (Fig. 8a).⁸⁹ Specifically, the hydroxyl group of CLA can be reacted with the isocyanate in a single step to form small molecular structures such as –CN, –CF₃, and –COO groups, while generating urea bonds to enhance Li⁺ conduction. Furthermore, the grafted CLA can be well-mixed with LATP to prepare a composite PEM, and the prepared CLA-CN-LATP PEM can exhibit an outstanding ionic conductivity of 1.25 × 10⁻³ S cm⁻¹ at 25 °C. Specifically, the grafted small molecule cyanide group can increase the interaction between Li⁺ and the polymer, favoring Li⁺ transport. In other words, the grafted small molecule cyanide group can attract and trap Li⁺, forming a complex of Li⁺ and cyanide. The presence of cyanide groups can enhance the interaction between the polymer chains and the ions, helping the ions to move through the electrolyte. The polar cyano-functional groups can attract Li⁺, thus forming an ion cloud around the polymer chain, making it easier for ions to move through the electrolyte and thus improving the ion transport properties of the electrolyte.

Fig. 8 Strategies for preparing cellulose-mediated PEMs with high ionic conductivity. (a) Schematic diagram of constructing a separator with high ionic conductivity in cellulose acetate grafted with ion-conducting molecules (reproduced with permission from ref. 89. Copyright 2023, Wiley-VCH). (b) Schematic illustration of Fe³⁺ coordinating with cellulose nanofibers and polyacrylic acid for Zn²⁺ transport, along with the ion conductivity of Fe³⁺-PAA/CNF (reproduced with permission from ref. 109. Copyright 2025, Wiley-VCH). (c) Schematic diagram of PEMs with high ionic conductivity prepared by compositing bacterial cellulose with an ion-conducting material. (reproduced with permission from ref. 110. Copyright 2022, Elsevier). (d) Schematic diagram of a separator with high ionic conductivity prepared by cross-linking bacterial cellulose with a multicomponent composite (reproduced with permission from ref. 111. Copyright 2025, American Chemical Society).

Crosslinking with metal-ions. As we mentioned above, long-range and ordered ion-conducting channels can be constructed between cellulose molecules using ionic cross-linking, and these long-range ordered ionic pathways are crucial for ion transport. Therefore, Chen and coworkers employed an Fe³⁺-carboxylate coordination strategy to synthesize a supramolecular hydrogel electrolyte exhibiting high ionic conductivity (32 mS cm⁻¹) (Fig. 8b).¹⁰⁹ Specifically, polyacrylic acid (PAA) and cellulose nanofibers (CNFs) rich in carboxyl groups were employed to form hydrogen bonds with Fe³⁺ and other supramolecular interactions. In particular, owing to the excellent structural stability exhibited by the Fe³⁺-PAA/CNF network in various solution environments, the Fe³⁺-PAA/CNF hydrogel demonstrates outstanding ionic conductivity stability, maintains flexibility, and retains exceptional ionic conductivity even at extremely low temperatures (1.8 mS⁻¹ at −40 °C). Moreover, leveraging the exceptional ionic conductivity of the Fe³⁺-PAA/CNF hydrogel and incorporating freeze-resistant mixed zinc salts, the resulting hydrogel electrolyte achieves ultra-high zinc cycle reversibility (averaging 99.4%), demonstrating outstanding cycling performance (a capacity retention of 81% and a specific capacity of 180 mAh g⁻¹) in aqueous zinc-ion batteries. Notably, during the coordination process with Fe³⁺, the resulting Fe³⁺-PAA/CNF hydrogel exhibits not only outstanding ionic conductivity but also demonstrates a synergistic enhancement in mechanical strength, flexibility, and adhesion, primarily due to the supramolecular forces generated by the coordination between Fe³⁺ and the PAA/CNF network.

Composite ion-conducting materials. To address the high crystallinity and low ionic conductivity of cellulose PEMs, Qi and coworkers prepared a cellulose-mediated PEM (r-CCE) with high ionic conductivity by compounding bacterial cellulose with Li_6.4La₃Zr_1.4Ta_0.6O₁₂ (LLZTO) (Fig. 8c).¹¹⁰ Notably, thanks to the decoupled segment structure of cellulose and the addition of ionic channels provided by LLZTO, r-CCE not only shows excellent ionic conductivity (1.68 × 10⁻³ S cm⁻¹), a significant Li⁺ transfer number (∼0.92), and a broad electrochemical window (∼5.3 V), but also contributes to the stabilization of the Li anode. In particular, the ionic conductivity of r-CCE gradually increases with the introduction of LLZTO. The ionic conductivity of r-CCE is three times that of the Celgard separator (0.36 × 10⁻³ S cm⁻¹) and BC separator (0.52 × 10⁻³ S cm⁻¹). In addition, the Li⁺ transfer number of r-CCE is 0.92, which is much higher than that of Celgard (0.39) and BC separator (0.42). Specifically, the Celgard separators have a relatively low porosity of about 45%, resulting in greater resistance to ion transport and lower ionic conductivity. As for the BC separator, its higher electrolyte adsorption capacity can facilitate electrolyte permeation, thereby reducing ion transport resistance. The bacterial cellulose chains can capture electrolyte solvent molecules by interacting with Li⁺ through polar groups, thus inhibiting the migration of PF6 anions to improve the number of Li⁺ transferred. For r-CCE, LLZTO, as a Lewis base, can greatly facilitate the dissociation of Li salts through strong binding to Li⁺, thus contributing to Li⁺ transport. From another perspective, LLZTO can provide additional high flux pathways for Li⁺ conduction, increasing the overall amount of Li⁺ conduction, thus facilitating Li⁺ homogenization and stabilization of Li deposition around the interface. Indeed, the higher Li⁺ conductivity and Li⁺ transfer number of the r-CCE compared to BC and Celgard are mainly due to the selective ion transport by LLZTO.

Composite multicomponent materials. From another perspective, excellent ion transport properties can also be achieved through composites between cellulose and other multi-component materials. In this regard, Yang and coworkers constructed a hydrogel electrolyte (IBVA) with high ionic conductivity for supercapacitors by introducing a bacterial cellulose (BC) network with weak ionic interactions (Fig. 8d).¹¹¹ Notably, the three-dimensional layered structure, abundant hydroxyl groups, high crystallinity, high Young's modulus, and high tensile strength make BC an ideal reinforcing element in a polymer matrix. Briefly, the IBVA PEMs with high ionic conductivity and mechanical properties were prepared by thermally initiated polymerization and salt solution immersion processes. BVA hydrogels were made by immersing BC membranes in PVA/AA precursors through a conventional gelation process. Furthermore, the prepared BVA hydrogels were immersed in 3 M ammonium formate (HCOONH₄) solution for “salting-in”, thereby reconstructing the polymer network and obtaining the IBVA PEM. In this process, H-BVA, C-BVA, and S-BVA were obtained by treating BVA hydrogels with different anions (HCOONH₄, NH₄Cl, and (NH₄)₂SO₄). In addition, formate ions can be used as intermediates to broaden the polymer network by weakly binding to cellulose chains, which further facilitates the migration of ammonium ions. The activation energies (E_a) for ion migration in H-BVA, C-BVA, and S-BVA were 0.18, 0.25, and 0.41 eV, respectively. Benefiting from the weak interactions between polymer chains and ions, the H-BVA PEM exhibited superior ionic conductivity (105 ± 5 mS cm⁻¹), which is much higher than that of IBVA PEM (10 ± 2 mS cm⁻¹) with (NH₄)₂SO₄. In particular, the ion conductivities of HCOONH₄, NH₄Cl, and (NH₄)₂SO₄ were 160 mS cm⁻¹, 175 mS cm⁻¹, and 208 mS cm⁻¹, respectively, suggesting that the salting out effect of the polymer network greatly limits the ionic transport efficiency of the liquid electrolyte.

4.2 Ion-selective enhancement

The ion selectivity of PEMs is a property that evaluates their ability to selectively transport specific ions or molecules.^112–115 For several energy-related devices (e.g., rechargeable batteries, fuel cells, reverse electrodialysis systems, etc.), the ion selectivity of PEMs takes on particular importance, as it forms the foundation for their successful operation. In this regard, cellulose can provide a new platform and pathway for the construction and design of novel ion-selective PEMs. Briefly, cellulose-mediated PEMs with high ion selectivity can be constructed through the following four strategies: (i) layered structure regulation; (ii) reverse charge shield; (iii) molecular sieve effect; and (iv) graded aperture engineering. This section will discuss the construction strategies and design principles for achieving selectivity in cellulose-mediated PEMs across various applications through specific examples.

Layered structure regulation. In nature, biological membrane systems can efficiently regulate the selective transport of ions across cell membranes through bio-ionic channels with unique structures and functions. However, biomembrane systems often exhibit poor mechanical properties, environmental stability, and operational performance, making it impossible to manufacture permeation energy devices using highly efficient biofilms. To develop cellulose-mediated PEMs with high selectivity for applications in osmotic energy conversion devices, Ye and coworkers developed a cellulose-mediated PEM with a layered structure that enhances selective energy harvesting by decoupling ion and electron transport (Fig. 9a).¹¹⁶ Specifically, negatively charged, highly oriented cellulose nanochannels and a conductive polyaniline network exist on both the inner and outer sides of the layered membrane, enabling highly efficient selective ion and electron transport. Moreover, the uneven distribution of cellulose on the polyaniline surface and the layered structure with carboxylation modifications were further confirmed by Raman imaging. Furthermore, the ion selectivity of the cellulose/PANI-based PEM was evaluated by calculating the cation migration number (t₊ < 1) and energy conversion efficiency (η < 50%). In brief, the t₊ and η of cellulose/PANI-based PEM were significantly higher than those of the blended membrane under a 50-fold concentration gradient. In the assembled reverse electrodialysis system, the cellulose/PANI-based PEM also exhibited higher power density and output performance compared to the blend membranes. That is, constructing multi-dimensional hierarchical structures within cellulose composites can facilitate the development of novel ion-electron decoupling pathways, enabling the creation of highly selective PEMs.

Fig. 9 Strategies for preparing cellulose-mediated PEMs with high ionic selectivity. (a) Schematic diagram for achieving high ion selectivity within a layered membrane and the cation migration number of layered/blend membranes at different KCl concentration gradients (reproduced with permission from ref. 116. Copyright 2024, American Chemical Society). (b) Schematic diagram of the structure of MoS₂ nanosheets/bacterial cellulose ion-selective membranes and molecular dynamics simulations of ion channels (reproduced with permission from ref. 117. Copyright 2024, The Royal Society of Chemistry). (c) Schematic diagram of ion-selective transport through CNF/MXene membranes at Zn anodes and the effect on coulombic efficiency (reproduced with permission from ref. 118. Copyright 2023, Wiley-VCH). (d) Schematic diagram of PCNH ion selectivity mediated by carboxymethyl cellulose and output voltage of PCNH-MEG (reproduced with permission from ref. 119. Copyright 2025, Springer Nature).

Reverse charge shield. The presence of channels with high charge selectivity and ionic conductivity is a prerequisite for electrolyte materials used in osmotic energy conversion. In this regard, harnessing osmotic energy via salinity gradients holds significant potential for sustainable power generation, which can be realized through 2D ion-selective nanofluidic devices. However, the traditional MoS₂ membrane suffers from low ion selectivity, resulting in reduced power generation efficiency. Therefore, Yang and coworkers proposed a reverse charge shield strategy to construct highly ion-selective channels within robust PEMs using bacterial nanocellulose (BC) and MoS₂ (Fig. 9b).¹¹⁷ Notably, the prepared MoS₂/BC PEMs showed an interlayer spacing of 9.8 Å, with the nanoscale channels demonstrating ideal negative charge density, thereby creating a favorable confinement environment: improving Na⁺ transport while blocking Cl⁻. BC is composed of β-D-glucopyranose units connected by β-1,4-glycosidic bonds and is an unbranched exopolysaccharide produced by certain bacteria, exhibiting a highly negative charge. Moreover, the design of MoS₂/BC PEMs incorporating BC and MoS₂ not only enhances mechanical properties but also improves ion-selective transport. In particular, MoS₂ can tightly bind to BC through hydroxyl bonds, while BC is confined within sub-nanoscale ion transport channels. The enhanced space negative potential effectively improves ion selectivity during permeation, achieving a power density of up to 73 W m⁻². To investigate the ion conduction mechanism within the PEM, the relationship between ion conductivity and ion concentration was demonstrated using KCl solutions of varying concentrations. In high-salt concentration regions, the ionic conductivity of MoS₂/BC PEM exhibited a linear relationship with KCl concentration, with values comparable to those of electrolyte solutions. In low-concentration regions, the ionic conductivity of MoS₂/BC PEM gradually stabilized and deviated from a linear relationship. Interestingly, an electric double layer forms when the surface charge of the MoS₂/BC PEM attracts counterions in the solution. The superior ion selectivity of MoS₂/BC PEM can be obtained when the dimensions of the nanofluidic channel are comparable to the thickness of the electric double layer in the PEM.

Molecular sieve effect. For alkali metal anode batteries, the uniform metal ion transport is a critical indicator for ensuring their durability, safety, and power density. In general, the decomposition of anions in the electrolyte leads to a reduction in metal cation concentration, which adversely affects the ionic conductivity, while the increase in local pH at the electrolyte/electrode interface further impedes Zn²⁺ transport during high-depth discharge. Therefore, Zhao and coworkers proposed a cellulose nanofiber/MXene membrane using molecular sieving to restrict anions and active water from the electrolyte/electrode interface through the dehydration effect of Zn²⁺, thereby achieving a stable Zn anode (Fig. 9c).¹¹⁸ Benefiting from the abundant functional groups of CNFs, the CNF/MXene coating achieves cation/anion selective molecular sieving while restricting water transmission, thereby effectively preventing corrosion during Zn stripping/plating processes. Therefore, the Zn anode coated with CNF/MXene exhibited long cycle life (∼3000 h), low voltage hysteresis, and excellent coulombic efficiency (99.7%) in Zn//Zn symmetric batteries. In particular, Zn//Zn symmetric batteries can operate at a high capacity of 100 mAh cm⁻² even at an extremely high current density of 100 mA cm⁻², corresponding to an exceptionally high Zn utilization rate of up to 88.2%. In other words, the CNF/MXene membranes with ion sieving properties can selectively remove anions and restrict the presence of active water at the electrolyte/electrode interface, achieving remarkable progress in regulating stripping and deposition processes, demonstrating significant application potential in other batteries. Notably, the ion selectivity of CNF/MXene was determined by building the current–voltage (I–V) curve under a Zn(CF₃SO₃)₂ electrolyte concentration gradient established across the membrane. By postulating that each ion species contributes a current given by the Nernst–Planck equation (parameterized by the effective diffusion coefficient D_i), the GHK model provides a method for quantitatively measuring selectivity, which proves useful for comparing the selectivity of different membranes (eqn (2)).where S_GHK is the selectivity ratio, c_low and c_high are the solution concentrations, k_b is the Boltzmann constant, e is the electron charge, and T is the temperature. Based on this equation, the selectivity ratio of Zn²⁺/CH₃SO₃⁻ in CNF/MXene was calculated to be 5, whereas glass fiber exhibited a selectivity of only 1, further demonstrating that CNF/MXene possesses outstanding cation/anion selectivity and prevents direct anion contact with the electrode/electrolyte interface.

Graded aperture engineering. Aperture engineering represents a highly promising strategy. Briefly, the graded aperture structure can significantly increase the specific surface area of the system in contact with surrounding ions/molecules, accelerating the formation of local ion/molecule concentration gradients. According to the Debye shielding effect, the electrical double layer overlaps within confined spaces when the aperture shrinks below the Debye length, causing the electric field strength produced by the surface charge on the aperture wall to increase exponentially. Notably, this intensified local electric field allows the surface charge density to govern ion selectivity. By precisely regulating the ion migration through the Stern layer, it establishes a mechanism for preferential counterion transport alongside virtually absolute co-ion exclusion. Thus, an aperture designed with dimensions smaller than the Debye length can dramatically increase selective ion transport behavior. In this regard, Qi and coworkers developed an in situ nanoconfinement strategy for preparing carboxymethyl cellulose (CMC) hydrogels with superior ion selectivity (Fig. 9d).¹¹⁹ The CMC was restricted within the delignified pomelo peel (DPP) framework, yielding a pomelo peel-confined CMC nanofluidic hydrogel (PCNH) material featuring uniformly distributed, high-charge-density, and stable nanopores with diameters smaller than the Debye length. Therefore, benefiting from the Debye shielding effect, the moist-electric generator (MEG) assembled based on PCNH exhibited superior ion-selective transport properties, with an open-circuit voltage about 0.4 V higher than CMCH without a graded aperture. Moreover, the PCNH-MEG can maintain peak performance at 1.48 V after continuous operation for 180 h, further demonstrating its exceptional operational stability.

4.3 Solubility resistance

For cellulose-mediated PEMs, their porous nanofiber networks enable efficient electrolyte penetration, high wettability to aqueous electrolytes, homogeneous ion transport paths, and good mechanical properties. Nevertheless, the contact with aqueous solutions can lead to the rearrangement of hydrogen bonds between cellulose molecules, leading to its swelling and deformation, thereby affecting ion transport channels.¹²⁰ Therefore, the solubility resistance/size stability of cellulose-mediated PEMs is another important indicator for evaluating their stability. In this regard, several strategies have been developed for enhancing the solubility resistance of cellulose-mediated PEMs, including: (i) ionic crosslinking/hydrogen bond shielding; (ii) multiple hydrogen bond mediating; (iii) strong hydrogen bond mediating; and (d) weak hydrogen bond mediating.¹²¹ This section will further summarize and discuss strategies and general design principles for achieving solubility resistance in cellulose-mediated PEMs with specific examples.

Ionic crosslinking and hydrogen bond shielding. Generally, the expansion of PEMs in aqueous electrolytes leads to uneven ion flux and irregular dendritic expansion. Thus, strategies are urgently needed to overcome swelling deformation while keeping the required porosity, to achieve widespread application of cellulose-mediated PEMs. In this regard, Ma and coworkers prepared a novel anti-swelling nanocellulose-based PEM (Zr-CNF) by in situ hydrolyzing carboxylated nanocellulose (TCNF) with Zr⁴⁺ to shield hydrogen bonds on the surface of nanocellulose and crosslink nanofibers (Fig. 10a).¹²² Notably, the hydrolysis process of Zr⁴⁺ can expand the pore size of the nanocellulose network, which can effectively inhibit the deformation and swelling of CNF in the aqueous electrolyte. For example, Zr-CNF and TCNF demonstrated different pore size changes after 3 days of immersion in 2 M ZnSO₄ aqueous solution. Compared to pristine TCNF, the thickness and pore size of Zr-CNF-based PEM did not change significantly before and after immersion, which implies that the Zr-CNF-based PEM has an excellent anti-swelling property, which is attributed to the coating of the Zr-ion hydrolysate and the interconnected crosslinked nanofiber structure. At a deeper level, the pore structure of the TCNF-based PEM is gradually deformed due to the interaction between water molecules and cellulose, which leads to ionic turbulence. In contrast, the Zr-CNF-based PEM can retain a homogeneous pore structure under the protection of Zr-hydrolysate coating, thereby achieving highly efficient and stable ion transport pathways. Therefore, with this excellent anti-swelling property, Zr-CNF-based PEMs exhibited superior cycle life (over 700 h at 10 mA cm⁻²) in assembled aqueous zinc batteries.

Fig. 10 Strategies for preparing cellulose-mediated PEMs with high solubility resistance. (a) Schematic illustration of the excellent solubility resistance of cellulose-mediated PEMs achieved by ionic cross-linking and hydrogen bond shielding (reproduced with permission from ref. 122. Copyright 2022, Wiley-VCH). (b) Schematic illustration of the excellent solubility resistance of cellulose-mediated PEMs through multilevel hydrogen bonding (reproduced with permission from ref. 123. Copyright 2024, Elsevier). (c) Schematic illustration of the excellent solubility resistance of cellulose-mediated PEMs achieved by strong hydrogen bonding (reproduced with permission from ref. 124. Copyright 2025, Elsevier). (d) Schematic illustration of the excellent solubility resistance of cellulose-mediated PEMs achieved by weak hydrogen bonding (reproduced with permission from ref. 111. Copyright 2025, American Chemical Society).

Multiple hydrogen bond mediating. To provide cellulose-mediated PEMs with excellent swelling resistance, Xu and coworkers dissolved cellulose in a solution containing Zn²⁺ and Al³⁺ and then co-polymerized acrylamide (AAm) and acrylic acid (AA) in it to construct a cellulose-mediated PEM with excellent anti-swelling properties (Fig. 10b).¹²³ Briefly, multiple hydrogen bonds are formed between the abundant –COOH groups, –NH₂ groups, and –OH groups belonging to AA, Aam, and cellulose, respectively. In the ZnCl₂/AlCl₃ solvent system, the hydrated Al³⁺ can react with –OH groups on cellulose to break the hydroxyl bonds between the cellulose chains. Then the hydrated Zn²⁺ disrupted additional hydrogen bonds, causing cellulose to dissolve at room temperature. Thanks to the multiple hydrogen bonds, the ion-C-P(AA-co-AAm) showed no significant change in volume after 14 days of immersion in deionized water, Na₂SO₄, and PBS buffer, which demonstrated its strong resistance to swelling (88.03%) and dimensional stability. Notably, cellulose serves a crucial role in providing –OH groups to form numerous hydrogen bonds.

Strong hydrogen bond mediating. Along similar design lines, Li and coworkers reported an anti-swelling PEM prepared by free radical polymerization under ambient conditions using the catechol redox reaction between phosphorylated lignocellulosic nanofibers and Ag⁺ (Fig. 10c).¹²⁴ In this process, numerous free radicals are generated within the system, which then initiates the polymerization of the monomers to form the polymer structure. Therefore, the resulting cellulose-mediated PEMs exhibited long-term water resistance due to the strong hydrogen bonding between the polymer chains that eliminated the interference of water molecules. According to the swelling curves, the PAGDP-Ag showed impressive insolubility characteristics with a value of 1.012 g g⁻¹ after 30 days of immersion in water, which was primarily attributed to the dense polymer network that reduced the surface energy by improving hydrogen bonding and therefore enhanced its resistance to swelling. Moreover, the formation of strong hydrogen bonds between DMSO and water molecules increased the amount of bound water, resulting in PAGDP-Ag with excellent resistance to freezing, i.e., good mechanical elasticity after 30 days at 20 °C. This work utilizes a strong hydrogen bonding network to provide a new design ideal for endowing cellulose-mediated PEMs with excellent resistance to swelling.

Weak hydrogen bond mediating. Some hydrogel-based PEMs typically exhibited greater solubility and weak mechanical strength in water/organic reagents due to the high water content and loose amorphous structure. Therefore, to change this dilemma and expand the application scenarios of cellulose-mediated PEMs, Zhang and coworkers constructed a cellulose-mediated PEM (IBVA) with excellent resistance to swelling by using bacterial cellulose (BC) and formic acid anions with a salting-out effect to build a weak hydrogen bonding network (Fig. 10d).¹¹¹ The “salt in” effect-based weak interactions of chaotropic formate anions were introduced into this system to demonstrate their positive effect on morphological stability. In brief, the IBVA-based PEM showed a uniform and compact lamellar structure with a spacing of 5 µm, with adjacent laminae interconnected to form a special hierarchical structure that incorporated the distinctive features of BC and PVA/PAA. Notably, this layered structure gives IBVA unique anti-swelling properties that effectively resist the absorption and swelling of ammonium formate solutions. Therefore, the maximum expansion of IBVA was 18%, while the expansion of pure PVA/PAA was 150%. This resistance to swelling can be attributed to the robust balance between permeability and elasticity within the IBVA network, which is determined by the weak hydrogen bonding and layered structure within the system.

4.4 Self-healing realization

Self-healing is a critical property of materials, enabling them to restore damaged areas through intrinsic, automatic repair mechanisms, thereby returning the material to its original condition, which helps maintain durability, safety, and reliability while minimizing sudden failures due to cracks or fractures.^125,126 In general, the self-healing ability depends on the reversibility of the cross-link bond, which could consist of dynamic covalent bonds, ionic bonds, hydrogen bonds, hydrophobic interactions, and supramolecular host–guest interactions.¹²⁷ The unique structural characteristics of cellulose enable cellulose-mediated PEMs to exhibit self-healing properties, primarily through mechanisms involving ionic bonding, hydrogen bonding, and melt-recrystallization transition.^128,129 Briefly, self-healing cellulose-mediated PEMs demonstrated prominent properties in applications such as rechargeable batteries, supercapacitors, and fuel cells. In this section, we analyze and discuss the strategies for the construction of self-healing cellulose-mediated PEMs through specific examples.

Non-covalent interactions: hydrogen and coordination bonds. To construct PEMs with high safety for flexible solid-state supercapacitors, Wen and coworkers proposed a cellulose nanofiber-reinforced polyacrylic acid/deep eutectic solvent PEM (PAA/DES/CNF) with excellent self-healing properties. In this process, the urea and ChCl in DES can act as hydrogen bond donors and hydrogen bond acceptors, respectively, providing a range of functional properties for the PAA/DES/CNF PEM. The doping of CNF can improve the mechanical properties of the PAA/DES/CNF PEM through two mechanisms: (i) the rigid structure of CNF fibers can provide mechanical reinforcement due to their crystalline regions; (ii) the abundant hydroxyl groups on CNF can establish hydrogen bonds with DES and PAA, forming a dynamic and dense hydrogen-bonding network. For the PAA/DES/CNF PEM, the homogeneous combination of various dynamic coordination bonds and hydrogen bonds in the polymer network results in excellent self-healing properties. In particular, these dynamic adhesives allow the PEM to spontaneously reorganize in the absence of external stimuli, thereby promoting the recovery of mechanical integrity. For example, the tensile properties of PEMs were almost completely restored after they were cut in half and self-healed for 24 h. Optical microscopy tests revealed that the healed samples showed only faint scars, indicating good self-healing. In brief, the dynamic interaction sites (dynamic coordination bonds and hydrogen bonds) within the PAA/DES/CNF PEM serve a key role in its self-healing behavior.

Following a similar research direction, Yang and coworkers proposed a molecular engineering strategy for designing a cellulose-based supramolecular zwitterionic hydrogel electrolyte (SZHE) with excellent self-healing capabilities for zinc ion capacitors (Fig. 11a).¹³⁰ Specifically, SZHE PEMs were prepared by introducing a recombinant network of non-covalent self-assembled supramolecular zwitterions into a cellulose nanofiber (CNF)-enhanced covalently cross-linked polyacrylic acid (PAA) network within ZnSO₄/EG electrolyte. Notably, the SZHE PEM can achieve excellent self-healing ability by reconstructing dynamic coordination bonds due to the uniform inclusion of numerous dynamic hydrogen bonds and coordination bonds in the SZHE hydrogel polymer network. In this regard, the scratches on the SZHE PEM became lighter after 24 h at room temperature and disappeared after 48 h. Therefore, zinc ion capacitors assembled based on self-healing SZHE PEMs demonstrate outstanding cycling stability and durability (2000 h), along with exceptional in situ repair capabilities, enabling the restoration of micropores induced by Zn-dendrites and cracks generated during plating/stripping processes. In other words, through the well-designed interaction force between multiple components, the prepared self-healing cellulose-mediated PEMs exhibit outstanding safety and durability in the assembled electrochemical devices.

Fig. 11 Strategies for preparing cellulose-mediated PEMs with self-healing properties. (a) Schematic diagram of the preparation process of supramolecular zwitterionic PEMs with self-healing properties (reproduced with permission from ref. 130. Copyright 2023, Royal Society of Chemistry). (b) Schematic diagram of the preparation process of the self-healing MXene/cellulose/BzMe₃NOH hydrogel through electrostatic effects (reproduced with permission from ref. 131. Copyright 2023, American Chemical Society). (c) Schematic diagram of the preparation process of CPPH with self-healing function through borate bonding (reproduced with permission from ref. 132. Copyright 2022, Wiley-VCH).

Non-covalent interactions: electrostatic effects. In addition to hydrogen and coordination-mediated self-healing, superior self-healing properties can be achieved by utilizing electrostatic effects between different molecules. In this regard, Tao and coworkers prepared a MXene/cellulose/benzyltrimethylammonium hydroxide (MCB) hydrogel with excellent self-healing ability by a solvent-assisted method (Fig. 11b).¹³¹ Briefly, MCB can be obtained by dissolving cellulose in benzyltrimethylammonium hydroxide (BzMe₃NOH) and mixing it with the MXene. Notably, the MXene creates dynamic cross-linking points between cellulose chains through electrostatic interaction, resulting in a dynamic network of MXene cross-linked cellulose chains, which is crucial for cellulose to exhibit self-healing properties. Specifically, the electrostatic interactions between MXene and terminal groups on the cellulose surface at the fracture reestablished new dynamic crosslinking points between the two materials when MCB was damaged, subsequently forming a novel 3D network structure. Moreover, the efficiency of self-healing in MCB hydrogels showed a decreasing trend with increasing cellulose and BzMe₃NOH content, a finding consistent with changes in chain flow relaxation time. Indeed, the mobility of cellulose chain segments within the system is a key factor in the strength of self-healing.

Covalent interactions: borate bonds. As a type of dynamic covalent bonding, borate bonds enable rapid healing of damaged materials at room temperature without external stimuli. Taking advantage of this feature, Nie and coworkers introduced a fast self-healing, stretchable cellulose-mediated PEM made by dynamic cross-linking via in situ polymerization of polyaniline (PANI) with TEMPO-oxidized cellulose (TOCNF) and PVA/borax (Fig. 11c).¹³² The TOCNF/PANI-PVA/borax (CPPH) exhibited superior stretchability and self-healing properties. This highly efficient self-healing capacity was attributed to the dynamic quadruple hydrogen bonding connections created between the tetrafunctional borate ions and hydroxyl groups in PVA, as well as the amine groups in TOCNF/PANI. In particular, the low-energy hydrogen bonds connecting TOCNF/PANI and PVA allow external forces to act preferentially on the breakage of the dynamic hydrogen bonds, resulting in a large polymer backbone extension that allows the material to withstand tensile strains of up to 1530%. That is, the self-healing property triggered by dynamic borate bonding exhibited outstanding performance. However, the self-healing materials based on borate bonds still suffer from the need for additional pH adjustment and weak mechanical strength.

4.5 Flame-retardant enhancement

Flame-retardant polymer electrolytes have become indispensable in improving the safety of rechargeable batteries and other energy-related devices.^133–135 In other words, with the increasing number of battery fires and explosions, it is particularly important to enhance the safety of rechargeable batteries from the perspective of materials design and selection. Particularly for cellulose materials, cellulose fibers are susceptible to pyrolysis when heated and the pyrolysis products are gaseous, liquid, and solid substances. Therefore, the inherent flammability and low oxygen index of cellulose impose higher demands on the safety performance of cellulose-mediated PEMs. In general, the flame-retardant properties of cellulose-mediated PEMs largely depend on the characteristics of the guest material. Therefore, this section primarily focuses on representative enhancement strategies for flame retardancy in cellulose-mediated PEMs.

Construct dynamic networks. Based on the need for the design of highly flame-retardant cellulose-mediated PEMs, Wang and coworkers proposed a strategy for the fabrication of highly flame-retardant cellulose-mediated PEMs, which is based on the injection of liquid PEG electrolytes into the flexible and porous CNF dynamics network (Fig. 12a).¹³⁶ Specifically, the dynamic anisotropic swelling was achieved by immersing the wetted CNF membranes prepared by vacuum filtration in NaCl solution, followed by protonation of the swollen gels at pH 2 to improve the stability of the PEMs. Furthermore, the CNF/PEG PEM was obtained by controlled progressive solvent exchange from water to ethanol to acetone followed by injection of a liquid PEG electrolyte consisting of PEG, succinonitrile (SCN), and lithium bis(trifluoromethanesulfonyl)imide (LiTFSI). In addition, PEG dimethyl ether with a molecular weight of 500 Da was used in battery testing to eliminate the adverse effects of PEG reactive and hydroxyl groups on battery performance. Notably, lithium iron phosphate (LFP) assembled based on CNF/PEGDME electrolytes exhibited superior safety. For example, the CNF/PEGDME electrolyte has good flexibility and the pouch cell will still work after at least 5 full folds and unfolds. In particular, flame testing clearly demonstrated that the CNF/PEGDME electrolyte has better flame retardancy than conventional liquid electrolyte-impregnated polyethylene spacers. Briefly, CNF/PEGDME electrolytes can retain their initial dimensions without shrinking and maintain their flame-retardant properties after exposure to internal flames for at least 50 seconds. In contrast, the polyethylene separator impregnated with propylene carbonate contracted for only 5 seconds after being ignited by the internal flame before catching fire. In other words, the mechanical flexibility and flame retardancy of CNF/PEGDME electrolytes make them exhibit outstanding safety in rechargeable batteries.

Fig. 12 Strategies for preparing cellulose-mediated PEMs with flame retardancy. (a) Schematic diagram of the preparation process and flame retardancy of CNF/PEG electrolyte (reproduced with permission from ref. 136. Copyright 2023, Wiley-VCH). (b) Schematic diagram of the flame-retardant mechanism of BC/Li-FR (reproduced with permission from ref. 137. Copyright 2025, Wiley-VCH). (c) Schematic diagram of DFM electrolyte with thermal shutdown characteristics and thermal treatment capability (reproduced with permission from ref. 138. Copyright 2024, Wiley-VCH). (d) Schematic diagram of the preparation process and flame retardancy of electrolyte prepared by surface modification of cellulose (reproduced with permission from ref. 139. Copyright 2024, American Chemical Society).

Composite flame-retardant materials. The current flame-retardant strategies (e.g., use of ceramic–polymer composites and nonflammable solvent formulations) tend to sacrifice electrochemical performance or scalability of PEMs and do not fundamentally address the trade-off between safety, environmental sustainability, and low-temperature functionality. In this regard, the development of novel flame-retardant materials for PEMs has also received increasing attention. Inspired by nature's evolutionary optimization of plant cell walls, Hu and coworkers fused bacterial cellulose with lignin-derived flame-retardant microspheres and Li-functionalized hydroxyapatite to prepare a nonflammable biomimetic cellulose-mediated PEM (Fig. 12b).¹³⁷ Notably, lignin-derived flame-retardant microspheres and Li-functionalized hydroxyapatite together are used as flame-retardant materials. Specifically, the innate char-forming ability of lignin combined with the fluorine/phosphorus-rich interface allows for self-extinguishing without affecting ion channels. Meanwhile, the inclusion of lithiated hydroxyapatite further introduces a bioinorganic interface to enhance thermal stability. Compared to conventional bacterial cellulose/LiTFSI, the bacterial cellulose/Li-FR can show excellent flame suppression with the synergistic effect of lignin and lithiated hydroxyapatite additives, with an 85% reduction in the peak heat release rate (298.74 kW m⁻²) and a significant reduction in the total heat release (7.95 MJ m⁻²). Moreover, the smoke generation analysis showed that the peak smoke release rate of bacterial cellulose/Li-FR was 38% lower than that of bacterial cellulose/LiTFSI, indicating more efficient smoke suppression. That is, flame-retardant properties can be effectively enhanced by adding flame-retardant materials to PEMs, which have shown excellent performance in various flame-retardancy tests.

Fast curing at high temperatures. For solid electrolyte rechargeable batteries, sharp metal dendrites may pierce the separator, leading to severe internal thermal runaway and short circuits. Moreover, batteries are susceptible to overheating under abnormal abuse conditions such as overcharging/discharging, ultra-high currents, and high/low operating temperatures due to unavoidable ohmic heating and electrochemical reaction heating. In particular, the separator inevitably shrinks and collapses extensively when temperatures exceed the melting point of the polyolefin separator, leading to severe internal short-circuiting and catastrophic thermal runaway. Therefore, to construct safer cellulose-mediated PEMs, Huang and coworkers proposed a safe deep-eutectic-polymer electrolyte (DFM) with a built-in thermal shutdown capability by taking advantage of the hydrophobic association of methylcellulose (MC) within a novel deep-eutectic-solvent (Fig. 12c).¹³⁸ Notably, methylcellulose chains can aggregate to form a dense polymer network due to hydrophobicity at high temperatures and break the solvation structure equilibrium within the deep-eutectic system by encapsulating Li⁺ in the polymer matrix, leading to rapid solidification of electrolyte. Furthermore, the temperature evolution in 1 Ah LiFePO₄ pouch batteries at 100% state of charge was investigated by an accelerated rate calorimeter test. Specifically, three critical temperatures were marked on the curve to analyze the thermal behavior of the battery, including the onset temperature of the self-heating process (T₁), the thermal runaway temperature (T₂), and the maximum temperature (T₃). In particular, the self-heating of LiFePO₄//LB-002//Li batteries started at 73 °C (T₁), whereas the DFM electrolyte can significantly increase the battery's T₁ to 172 °C. In the case of DFM-based batteries, a highly stabilized solid electrolyte interface can slow the self-heating process by eliminating the exothermic reaction that typically occurs with carbonate solvents at elevated temperatures and reducing the production of flammable gases. Notably, the hydrophobic interactions of MC substituents increase at temperatures above 80 °C, causing the MC chains to aggregate to form a dense polymer network, leading to electrolyte solidification. Therefore, the synergistic effects of MC hydrophobic association and solvation structural changes lead to rapid solidification of the electrolyte at elevated temperatures, which completely cuts off Li⁺ transport and terminates the battery operation. That is, this work proposed a smart cellulose-mediated PEM that can spontaneously solidify based on temperature increase, leading to a high level of battery safety.

Surface coating. For lithium-ion batteries, the hydroxyl groups on the surface of the cellulose gel polymer electrolyte can side-react with the lithium, resulting in poor interfacial compatibility and low safety. In this regard, Guo and coworkers fabricated a novel cellulose-mediated PEM with good interfacial compatibility and excellent flame retardancy by uniformly coating nanohydrotalcite/PVDF-HFP composites on the surface of cellulose membranes (nanohydrotalcite/PVDF-HFP@cellulose@nanohydrotalcite/PVDF-HFP) using electrospinning technology (Fig. 12d).¹³⁹ Notably, the nanohydrotalcite/PVDF-HFP layer on the surface of the cellulose membrane can effectively prevent the hydroxyl groups on its surface from coming into contact with the lithium, thus alleviating the side reactions and improving the interfacial compatibility. Meanwhile, the nanohydrotalcite layer can act as a redistributor of Li⁺ transport during the passage of Li⁺, which ensures uniform deposition of Li⁺ and thus reduces the formation of Li dendrites, which can significantly enhance the cycling stability of cellulose gel polymers. The composite membranes of PCP-2%, PCP-4%, and PCP-6% were obtained by adding 2%, 4%, and 6% nanohydrotalcite in PVDF-HFP solution, respectively. The cellulose membrane and PVDF-HFP@cellulose@PVDF-HFP (PCP) membrane continue to burn after the flame has left, whereas 2%, 4%, and 6% PCP membranes do not continue to burn, which is mainly due to the excellent flame-retardant properties of nanohydrotalcite. Specifically, the layered hydroxyl groups and interlayer ions of hydrotalcite will release water and carbon dioxide at high temperatures, resulting in structural changes that reduce the concentration of combustion gases, block O₂, and thus act as a flame retardant.

4.6 Porosity optimization

The pore characteristics (e.g., pore size distribution, fiber width, fiber length, permeability, porosity) of PEMs are closely related to the overall performance (e.g., electrolyte absorption and elastic modulus) in electrochemical devices and cell performance.¹⁴⁰ In general, the smaller pore size and larger porosity of the separator are favorable to the performance of the battery, which means greater ionic conductivity and higher insulation safety. For cellulose-mediated PEMs, the key feature lies in the porosity optimization that can be achieved for PEMs. It is worth noting that the porosity of PEMs is primarily controlled by the preparation method, so this section will mainly focus on the preparation methods of cellulose-mediated PEMs with different porosities.

Solution-phase inversion. Indeed, PEMs with higher porosity can ensure superior electrolyte retention and greater ionic conductivity, thereby reducing heat generation and energy loss. However, the complex preparation process for cellulose-mediated PEMs has somewhat limited their utilization in large-scale energy storage applications to some extent. Therefore, Yu and coworkers prepared a transparent, flexible, and reproducible mesoporous cellulose-mediated PEM (mCel) through a simple and scalable solution-phase inversion process using ionic liquids (ILs) as solvents (Fig. 13a).¹⁴¹ Briefly, cellulose can be easily dissolved in 1-butyl-3-methylimidazolium chloride ([Bmim]Cl) and the PEM is subsequently regenerated using pure water as a non-solvent. In this process, the excellent solubility of cellulose in [Bmim]Cl was ascribed to the hydrogen bonding interactions formed between the chloride anion/imidazole cation in [Bmim]Cl and the hydroxyl protons of cellulose. Furthermore, water gradually replaced [Bmim]Cl when the Cel/IL gel was immersed in distilled water, diffusing from the gel surface into the gel body. This spontaneous molecular self-assembly process leads to the rearrangement of cellulose chains to form swollen regenerated cellulose-mediated PEMs. Notably, the porosity of mCel-PEM was 71.78%, which was higher than cellulose separator paper (NKK-TF4030, 29.68%) and the nonwoven polypropylene membrane (NKK-MPF30AC, 51.08%). The high porosity offers abundant ion diffusion pathways and provides ample space for electrolyte adsorption. Therefore, the highly flexible mCel-PEM-based supercapacitor showed a high volumetric capacitance of 191.66 F cm⁻³ and areal capacitance of 153.34 mF cm⁻² at 10 mV s⁻¹.

Fig. 13 Strategies for preparing cellulose-mediated PEMs with high porosity. (a) Schematic diagram of high porosity cellulose-mediated PEMs prepared by dissolving cellulose via solution phase transition (reproduced with permission from ref. 141. Copyright 2017, Wiley-VCH). (b) Schematic diagram of the preparation of high porosity separators by organic solvent modulated cellulose molecular reassembly (reproduced with permission from ref. 142. Copyright 2025, Springer Nature). (c) Schematic diagram of the preparation of high porosity separators by freeze-drying and hot-pressing cellulose (reproduced with permission from ref. 143. Copyright 2025, Elsevier). (d) Schematic diagram of the preparation of a hierarchical porous cellulose separator using vacuum filtration (reproduced with permission from ref. 144. Copyright 2022, Wiley-VCH).

Cellulose molecular reassembly. As we mentioned above, the abundance of hydroxyl groups on cellulose surface leads to strong hydrogen bonding between fibers, forming a nonporous and dense membrane, which results in excellent mechanical properties at the expense of Li⁺ transport. In this regard, employing cellulose molecular reassembly, freeze drying, and vacuum-filtration has been shown to obtain suitable porosity. For example, Xu and coworkers designed a highly porous regenerative cellulose separator (SORC) for supercapacitors (Fig. 13b).¹⁴² In this process, efficient solubilization and rapid regeneration of cellulose were achieved using superbase-derived ionic liquids and DMSO, which facilitated the formation of recombinant nano-cracked membranes of cellulose molecules. Specifically, these unique nano-cracks with an average width of 7.45 nm were created by DMSO modulating hydrogen bonding to accelerate cellulose molecular recombination, which endowed the separator with excellent electrolyte uptake (329%) and high porosity (70.2%). Indeed, the high porosity of PEMs can offer numerous ion diffusion channels and sufficient space for electrolyte uptake. Compared to the dense cellulose membrane (SRC, 44% and 254%), nonwoven polypropylene membrane (MPF30AC, 37.1% and 164%), and cellulosic paper (TF4030, 63.4% and 297%), the prepared SORC membranes demonstrated significantly enhanced porosity and electrolyte absorption. Therefore, supercapacitors assembled with SORC achieved an outstanding capacitance retention of 99.5% after 10 000 cycles and a high specific capacitance of 93.6 F g⁻¹ at 1.0 A g⁻¹.

Freeze drying or vacuum filtration. Along different design directions, Deng and coworkers prepared a lithium-ion battery separator (BHM) with excellent porosity from bacterial cellulose (BC), nanohydroxyapatite (HAP), and melamine polyphosphate (MPP) by freeze-drying (Fig. 13c).¹⁴³ Notably, the porosity of BHM (>70%) is nearly doubled compared to the Celgard 2325 separator (36.17%), which is ascribed to the formation of a porous structure during the freeze-drying process. In other words, the uniform and high-density pore size distribution of the separator can effectively facilitate ion transport. Moreover, the affinity of the polar groups of BC also contributes to the uptake of electrolyte, while the excellent electrolyte wettability enhances the affinity between the electrolyte and separator, thereby improving ionic conductivity. Moreover, Zhang and coworkers reported a hierarchical porous cellulose separator (HPC) with high porosity for potassium batteries by engineering the natural supramolecular structure of cellulose (Fig. 13d).¹⁴⁴ Specifically, ramie was chosen as the raw material and was decomposed into cellulose nanofibers by TEMPO oxidation and strong agitation. The HPC separator was prepared by filtering the obtained cellulose nanofiber solution. Notably, the supramolecular and hierarchical structure of cellulose nanofibers gives the HPC high porosity, excellent mechanical properties, and superior electrolyte wettability. Specifically, matrix removal of lignocellulose and nano-fibrillation of cellulose introduced pores and channels of different sizes in the supramolecular structure. The cellulose suspension was washed with deionized water and then re-dispersed in ethanol, where the space between the fibers was improved due to the solvation effect, with hydrogen bonding interactions in ethanol being weaker than that in water. In particular, the rich hierarchical interactions and spatial structure between cellulose nanofibers enable the HPC separator (electrolyte uptake, 324%) to have a higher electrolyte uptake than the GF separator (electrolyte uptake, 303%). Therefore, potassium ion batteries assembled based on HPC separators with enhanced electrolyte uptake exhibited superior cycling performance and excellent rate capability.

4.7 Mechanical reinforcement

Cellulose is a linear natural polymer compound composed of β-D-glucopyranose units interconnected by repeating units of cellobiose. Indeed, the existence of sugar bonds decreases the hydrolysis stability of cellulose macromolecules to some extent. Specifically, reactions with water under acid/alkali conditions and elevated temperature conditions cause the breakdown of glycosidic bonds and degradation of cellulose macromolecules, resulting in reduced thermal stability and mechanical properties of cellulose fibers. Therefore, the issue of pure cellulose exhibiting low mechanical strength under high temperatures and high humidity conditions has long been a bottleneck restricting its application in battery separator applications.^145–147 For PEM-mediated energy devices, the mechanical properties of PEMs profoundly affect their safety. In this section, we focus on cellulose-mediated PEMs with the aim of analyzing and discussing strategies and methods for the construction of high mechanical strength cellulose-mediated PEMs.

Molecular-crosslinking. To construct cellulose-mediated solid electrolytes with high mechanical strength, Zhou and coworkers synthesized a hydrogel electrolyte featuring a physically crosslinked network using polyoxometalate (silicotungstic zinc, STA)-crosslinked cellulose with a Keggin structure, which exhibited outstanding mechanical properties (a tensile strength exceeding 3 MPa, a toughness of 1.5 MJ m⁻³, and an elongation at break over 80%) (Fig. 14a).¹⁴⁸ Briefly, cellulose was dissolved in a neutral pH binary solvent system (TMG/MAA, mass ratio 1:1), followed by the addition of a small amount of hydrated tungstosilicic acid to crosslink the cellulose molecular chains, and finally regenerated using water as a solidification bath to obtain the hydrogel. In this process, the highly polarized metal–oxygen bonds in polyoxometalates interact with cellulose hydroxyl groups through a combination of hydrogen bonding and coordination interactions, significantly enhancing the mechanical strength of the resulting cellulose hydrogel. In particular, this mechanical toughness not only ensures structural integrity under mechanical deformation but also maintains stable ion transport under stress, which is crucial for achieving long-term cycling stability and practical applications in wearable flexible devices. It is noteworthy that during the crosslinking process between STA and cellulose, the mechanical strength of the resulting hydrogel electrolyte depends on hydrogen bonding interactions between cellulose chains and supramolecular interactions between STA and cellulose. Increasing cellulose content leads to enhanced tensile strength and elongation, attributable to a more tightly packed hydrogen bond network. However, for cellulose-mediated PEMs, pursuing high mechanical strength is not always advisable, as high mechanical strength typically compromises ionic conductivity. Therefore, a balance must be struck between mechanical strength and ionic conductivity.

Fig. 14 Strategies for preparing cellulose-mediated PEMs with high mechanical properties. (a) Schematic diagram of preparing high-strength gel electrolytes by crosslinking cellulose with polyoxometalate molecules (reproduced with permission from ref. 148. Copyright 2025, Wiley-VCH). (b) Schematic representation of a cellulose-mediated PEM with high mechanical strength prepared by graft modification and solution recasting (reproduced with permission from ref. 88. Copyright 2024, Springer Nature). (c) Schematic representation of a cellulose-mediated PEM with high mechanical strength prepared by structural transformation induced by alkali treatment (reproduced with permission from ref. 98. Copyright 2023, Wiley-VCH).

Graft modification. Along different design lines, Cao and coworkers demonstrated a molecular engineering strategy to convert inert cellulose into PEMs with high mechanical strength and ionic conductivity properties (Fig. 14b).⁸⁸ Specifically, cellulose was converted into cellulose phthalate (CP) with excellent mechanical properties through a homogeneous esterification reaction in which phthalate anhydride replaces the hydroxyl groups of cellulose. Moreover, the introduced carboxyl groups can form a strong hydrogen bonding network between the CP chains, which can enhance the mechanical strength of the membrane even at high lithium salt content. Notably, the CP-based PEM was prepared by mixing the obtained CP with lithium salt and using ethyl carbonate as a plasticizer by solution-casting. In addition, CP-based PEMs also have excellent tensile strength (12 MPa) with an elongation of 80%, indicating that plasticizers positively influence the mechanical properties of cellulose-mediated PEMs. Comparatively, the separator based on cellulose benzoate and cellulose acetate exhibited low mechanical strength with tensile strengths of 0.32 MPa and 0.22 MPa, which can be explained by the strength of the hydrogen bond network. In particular, the prepared CP separator has shown significant superiority in realizing high ionic conductivity and high mechanical strength, a trade-off in properties that has long been a bottleneck limiting separator development.

Structural transformation. In addition, to design fully cellulose-mediated PEMs with high stability, Hu and coworkers proposed a nanocellulose–carboxymethylcellulose (CMC) electrolyte with high mechanical strength for an aqueous zinc ion battery (Fig. 14c).⁹⁸ The cellulose–CMC was prepared using vacuum filtration and NaOH treatment. Notably, NaOH treatment greatly increased the wet-state mechanical strength of the cellulose–CMC composite membrane, which is essential to resist perforation by zinc dendrites. In particular, the tensile strength of NaOH-treated cellulose–CMC membranes in the wet state was 72 MPa, which was much higher than that of cellulose–CMC membranes before NaOH treatment and glass fiber separators. The enhancement in mechanical strength stems from the transformation of cellulose's molecular structure during NaOH treatment. Specifically, the pristine cellulose and cellulose–CMC membranes showed cellulose I structure before NaOH treatment, and the cellulose I structure changed to the cellulose II structure after NaOH treatment. In comparison, the cellulose II structure has a higher tensile strength than the cellulose I structure. That is, the NaOH treatment can help to increase the mechanical strength of cellulose membranes without altering the functional group of the bio-polymer. Briefly, this work obtained cellulose-mediated PEMs with outstanding mechanical properties using alkali treatment, demonstrating the effect of structural transformation of cellulose after alkali treatment on the mechanical properties. In particular, the low-cost, sustainability, high mechanical strength, and ease of fabrication of cellulose–CMC PEMs give them significant potential in zinc ion batteries, paving the way for grid-scale energy storage powered by renewable energy.

Indeed, the mechanical strength of PEMs for rechargeable batteries has been the focus of researchers’ attention. Too high or too low mechanical strength can significantly impact the electrochemical performance of the devices. In particular, controlling the balance between the mechanical strength and other properties of PEMs is a top priority. In this regard, exploring simple and scalable ways to improve the mechanical properties of cellulose-mediated PEMs without compromising other properties requires special attention, which may be key to achieving their high stability.

4.8 Photoelectric stability enhancement

Despite the enormous potential for commercial applications, photoelectric conversion devices represented by perovskite solar cells still face challenges in long-term stability.^149,150 Briefly, polycrystalline perovskite films processed by solution methods inevitably develop high-density defects at their surfaces and grain boundaries, which serve as charge traps to exacerbate ion migration and non-radiative recombination, severely compromising device stability and efficiency.^151,152 In this regard, various strategies have been developed to overcome these challenges, such as interface engineering, defect passivation, and chiral nematic structure preparation. Typically, organic small molecules and polymers with different functional groups can be employed to optimize interfacial contact at perovskite grain boundaries and at the perovskite/hole-transporting materials. Compared to small molecules, polymers exhibit higher stability in photoelectric devices due to their abundant functional group sites and long-chain molecular structure. In this regard, cellulose derivatives such as cellulose acetate butyrate, cellulose acetate, ethyl cellulose, and hydroxyalkyl cellulose can effectively enhance device stability when incorporated into perovskite solar cells due to the presence of lone pairs on O atoms and abundant hydrogen bonding interactions.^153–155 Therefore, this section focuses on cellulose's role in photoelectric conversion devices, comprehensively analyzing its function in enhancing conversion stability and construction strategies.

Interface engineering. To develop green, low-cost, and efficient biomass interfacial materials for photoelectric devices, Chen and coworkers prepared a cellulose derivative (6-O-[4-(9H-carbazol-9-yl)butyl]-2,3-di-O-methyl cellulose, C-Cz) as a dual-functional PEM between the perovskite layer and the hole transport layer to enhance photoelectric stability (Fig. 15a).¹⁵⁶ Specifically, C-Cz was prepared by introducing methyl groups at the O-2 and O-3 positions, and a carbazole group at the O-6 position of cellulose through a four-step reaction (yield: 73%). Notably, the prepared C-Cz exhibited superior energy alignment, outstanding thermal stability, and strong interaction with the perovskite surface, which is crucial for carrier transport. Moreover, the methoxy groups and the resulting hydrogen bonds promote the passivation of surface defects in the perovskite, effectively inhibiting carrier complexation. The electrostatic charge distribution on C-Cz was investigated by calculating the electrostatic surface potential (ESP). Partial negative charges are delocalized on the O atoms of ether groups and the methoxy, which can passivate Pb²⁺ with insufficient coordination on the perovskite surface. Therefore, owing to the excellent structural stability of C-Cz and its strong interfacial interaction with the perovskite surface, devices based on C-Cz maintained over 88% of their original photovoltaic conversion efficiency after 2800 h. That is, cellulose derivatives prepared using molecular engineering strategies demonstrate unique application potential in enhancing the photoelectric stability of photoelectric conversion devices.

Fig. 15 Strategies for preparing cellulose-mediated PEMs with photoelectric stability. (a) Schematic diagram of the interaction between perovskite and C-Cz, as well as the PCE of the PSCs with and without C-Cz (reproduced with permission from ref. 156. Copyright 2023, Wiley-VCH). (b) Chemical structures of C-p-Py and C-m-Py as well as the PCE of unencapsulated PSCs with C-p-Py and C-m-Py (reproduced with permission from ref. 157. Copyright 2025, Springer Nature). (c) Schematic diagram of the preparation of the CNC/PVA gel and the optoelectronic properties (reproduced with permission from ref. 158. Copyright 2025, American Chemical Society).

Passivation defects. Defect passivation is an effective method for optimizing the interface between the perovskite and the electron transport layer in solar cells. To enhance the photoelectric stability of solar cells, Gao and coworkers prepared esterified cellulose (C-p-Py and C-m-Py) by introducing different N-position pyridines into cellulose molecular chains as an electron-selective electrolyte interface material between SnO₂ and the perovskite (Fig. 15b).¹⁵⁷ Notably, C-m-Py with a meta N atom exhibited a higher dipole moment and a more closely matched energy arrangement compared to C-p-Py. In particular, C-m-Py can concurrently improve the conductivity of SnO₂ and its passivation capability for perovskites, thereby improving interfacial charge extraction efficiency. Briefly, the synthesis of C-m-Py and C-p-Py involved the uniform one-pot modification of cellulose using 3-(pyridine-3-yl)acrylic acid and 3-(pyridine-4-yl)acrylic acid in DMAc/LiCl, respectively. This homogeneous one-pot modification strategy for cellulose does not cause significant pH fluctuations during the conversion process, thereby ensuring reduced degradation of cellulose chains. C-p-Py and C-m-Py exhibited excellent solubility in both dimethyl sulfoxide and N,N-dimethylformamide, ensuring superior solution processing properties. Moreover, compared to pure SnO₂ (22.71%), the efficiencies of perovskite solar cells treated with C-m-Py and C-p-Py were significantly enhanced to 23.79% and 24.35%, respectively. After 2500 h of storage, the photoelectric conversion efficiency of devices treated with C-p-Py and C-m-Py remained above 85% of their initial efficiency. This study reveals an effective strategy for designing low-cost interface materials for solar cells using cellulose and elucidates the modulation of functional groups through molecular engineering of cellulose to enhance materials's electronic properties and interfacial characteristics.

Chiral nematic structure. Mechanically responsive materials visualize invisible stresses, triggering immediate and intuitive visual feedback to provide instant and effective indications of dynamic physical stimuli. Structural colors generated by photonic crystals offer a robust coloring method, arising from the interaction between the mesoscale periodic structures of photonic crystals and light of specific wavelengths. In this regard, cellulose nanocrystals (CNCs) are regarded as highly promising sustainable and biocompatible building blocks for constructing photoelectrically responsive polymer electrolyte materials due to their ability to self-assemble into chiral nematic structures. However, the inherent thinness and structural sensitivity of CNC membranes may limit the incorporation of additional components, thereby restricting the enhancement of supplementary properties. Therefore, Fang and coworkers developed a CNC-based mechanically color-changing conductive electrolyte material (CNC@PVA) with dual-signal responsiveness featuring interactive photoelectric feedback (Fig. 15c).¹⁵⁸ The color-developing component of CNC/PVA is attributable to its adjustable CNC structure, while its conductivity stems from its inherent ion migration. Specifically, the mechanical response was achieved through the integrated processes of polyvinyl alcohol (PVA) encapsulation, interpenetration, and crystallization, including deformation in the encapsulated CNC photonic crystal core and thereby triggering dynamic changes in conductivity and color. Interestingly, the prepared CNC@PVA composite electrolyte materials exhibited dual capabilities for dynamic optical strain detection and electrical signal conduction, demonstrating remarkable potential in human health monitoring.

Although cellulose-mediated PEMs have achieved diverse properties, the following key aspects still need to be addressed: (i) the balance between performance and structural stability: to achieve high ionic conductivity in cellulose-mediated PEMs, substantial amounts of flexible chains and ionic groups need to be introduced, which may cause swelling and softening of the polymer network, resulting in a sacrifice of mechanical strength. Moreover, the precision and stability of porosity demand particular attention. (ii) Interface interaction with the electrode: evaluation of interfacial wettability, chemical stability, and dendrite suppression performance of cellulose-mediated PEMs at electrode interfaces. (iii) Reliability verification in actual operating environments: laboratory testing is typically conducted under mild conditions (low current and limited cycles), but practical applications require evaluating performance degradation under high loads, across a wide temperature range (−20 °C to 60 °C), and with higher cycle counts. Concurrently, key safety parameters such as the thermal runaway trigger temperature, heat generation rate, and gas composition of cellulose-mediated PEMs must be evaluated.

5. Applications of cellulose-mediated PEMs

As we mentioned above, cellulose-mediated PEMs can be endowed with special properties, such as superior ionic conductivity, high porosity, high mechanical property, superior ion selectivity, outstanding flame retardancy, self-healing properties, and prominent photoelectric stability with the aid of precise engineering strategies. Given these compelling properties, cellulose-mediated PEMs have yielded fruitful research outcomes across diverse applications, with their use in energy storage gaining increasing attention. Particularly, the emergence of cellulose-mediated PEMs has significantly broadened the forms, applications scenarios, and properties of green renewable materials for energy storage applications, such as rechargeable batteries, capacitors, fuel cells, solar cells, and reverse electrodialysis systems (Fig. 16). In this section, we critically discuss the applications of cellulose-mediated PEMs in energy storage through the latest and most representative examples.

Fig. 16 Overview of key properties and applications of cellulose-mediated PEMs.

5.1 Rechargeable lithium batteries

Solid-state batteries utilizing lithium metal anodes represent the next generation of energy storage systems, offering high energy density and enhanced safety. Indeed, the implementation of such batteries will primarily depend on the development of novel PEMs and the construction of ionic conduction networks within cathode materials.^159–161 Typically, poly(ethylene oxide) (PEO) is a common candidate for PEMs because of its low density, ease of processing, ability to dissociate lithium salts at high temperatures, and superior interfacial contact with the electrode.¹⁶² However, the Li⁺ transport in PEMs is strongly coupled with segmental motion of polymer chains, leading to limited ionic conductivity (generally less than 10⁻⁵ S cm⁻¹) and a low Li⁺ transfer number (0.2–0.5). Therefore, as we mentioned above, to construct green and efficient cellulose-mediated PEMs, Hu and coworkers fabricated a high-performance cellulose nanofiber (CNF)-based solid polymer ionic conductor (Li–Cu–CNF) by coordination of copper ions with 1D cellulose nanofibers.⁹⁴ The Li–Cu–CNF ionic conductor was fabricated through simple ion coordination and solvent exchange processes. Briefly, the CNF was first immersed in a Cu²⁺ saturated alkaline solution, followed by washing off NaOH with water and replacing the water with DMF, and then immersed in LiPF₆ electrolyte solution to facilitate the formation of the Li–Cu–CNF ionic conductor. Therefore, benefiting from the cross-linking of Cu²⁺ with cellulose molecular chains, the prepared Li–Cu–CNF material exhibited excellent Li⁺ conductivity (1.5 × 10⁻³ S cm⁻¹ at 25 °C) and electrochemical properties. In particular, the Li–Cu–CNF electrolyte can maintain stable Li cycling performance in a Li symmetric battery for 300 h at a current of 0.5 mA cm⁻² without dendrite-induced short circuits.

Moreover, in addition to lithium-ion batteries, the composite PEMs prepared from bacterial cellulose have shown excellent performance in lithium–sulfur batteries. In this regard, Guo and coworkers fabricated a cellulose-based Janus PEM (Janus-S) with superior mechanical/thermal properties and superior interfacial functionality using interfacial engineering strategies to achieve long-cycle-life lithium-sulfur batteries (LSBs) (Fig. 17a).¹⁶³ Briefly, CNF polymer chains bearing oxygen-containing functional groups exhibit strong lithium affinity at the anode interface, promoting uniform Li⁺ flux formation through self-polymerization effects. At the interface between the cathode and PEMs, the uniformly exposed single-atom Ru on the reduced graphene oxide surface can “capture” polysulfides, thereby lowering their activation energy and enhancing the conversion kinetics rate. Notably, the Janus-S PEMs with a well-developed pore structure were prepared under vacuum-assisted conditions by sequential addition of a blended suspension to a filter membrane, followed by a freeze-drying process. In addition, suppressing Li dendrites is critically important for the practical application of lithium–sulfur batteries. In a typical liquid electrolyte, Li⁺ near the surface of the Li anode was reduced due to slower Li⁺ migration, resulting in a Li⁺ depletion layer. For the boron nitride nanosheets/CNF composite layer (BNNs@CNF), the self-concentration adsorption of Li⁺ on the polymer chains can significantly prolong the Sand's time and promote initial uniform nucleation. With these properties, the Li//BNNs@CNF//Li symmetric cell showed a low voltage hysteresis of about 20 mV at an average capacity of 2 mAh cm⁻² and can operate stably for 200 h at 2 mA cm⁻². These results indicated that the functional BNNs@CNF layers had great potential to fulfill the high charge/discharge rate requirement of practical LSBs.

Fig. 17 Application of cellulose-mediated PEMs in rechargeable batteries. (a) Schematic diagram of a cellulose-based Janus PEM with superior mechanical/thermal properties for lithium–sulfur batteries (reproduced with permission from ref. 163. Copyright 2022, Wiley-VCH). (b) Schematic diagram of a Zn/I₂ battery with an integrated bacterial cellulose-mediated PEM/iodine cathode and the cycling stability of the Zn//Zn symmetric battery with GF and BC (reproduced with permission from ref. 70. Copyright 2025, Royal Society of Chemistry). (c) Schematic diagram of an oxidized cellulose membrane as a sodium ion battery separator and cycle properties of an Na//Na symmetric battery (reproduced with permission from ref. 179. Copyright 2025, Wiley-VCH). (d) Schematic diagram of the preparation process and function of CCP@TiO₂ and the cycling stability of Al//Al symmetric batteries with GF, CCP, and CCP@TiO₂ at 0.5 mA cm⁻² and 1 mAh cm⁻² (reproduced with permission from ref. 187. Copyright 2024, Elsevier).

5.2 Rechargeable zinc batteries

Among various energy storage systems, rechargeable zinc batteries are of interest due to their high safety, low price, and low redox potential and high theoretical capacity of the zinc anode.^164–166 However, the inherent rough surface of metallic zinc causes uneven distribution of electric/ionic fields on the zinc anode, generating heterogeneous zinc deposits that gradually evolve into zinc dendrites during prolonged charging/discharging.^167–169 The zinc dendrites may puncture the PEMs and eventually short-circuit the zinc cell. Therefore, the construction of suitable PEMs to inhibit zinc dendrites has been widely investigated. In this regard, Zhang and coworkers assembled a robust 2D MXene-based hydrogel film crosslinked by a 1D cellulose nanofiber dual network as an interfacial layer to stabilize the zinc anode (Fig. 17b).¹⁷⁰ The MXene–CNF composite membrane was assembled by vacuum-assisted filtration using a mixed colloidal dispersion of MXene and CNF. The obtained dry MXene–CNF membranes exhibited superior mechanical integrity and distinct metallic luster and showed lamellar microstructures with a thickness of about 8 µm. Furthermore, the MXene–CNF membranes were immersed in an aqueous solution of 2M ZnSO₄ to absorb a significant amount of the solution into a hydrogel with a liquid content of up to 60 wt%, which can be attributed to the strong water-absorbing and water-retaining capacity of the reinforced CNF network in the MXene–CNF membranes. To further understand the zinc nucleation and deposition behavior on different zinc anode surfaces, theoretical simulations using finite element calculations were carried out to reveal the effects of the MXene–CNF hydrogel interfacial layer on the electric field distribution, the Zn²⁺ concentration field and the stress–strain field. Compared to the bare Zn anode, the Zn anode in the MXene–CNF hydrogel interface exhibited a uniform electric field strength distribution, which greatly facilitated the reduction of Zn²⁺ aggregation and decreased the possibility of dendrite formation. Moreover, the MXene–CNF/Zn anode presented a more uniform Zn²⁺ concentration field near the electrode surface, which provided a more stable supply of Zn²⁺ and thus promoted more uniform Zn deposition. Indeed, these results were attributed to the MXene–CNF hydrogel interfacial layer having abundant zincophilic sites and low Zn²⁺ migration barriers. With these properties, the Zn/MXene–40%CNF/Zn symmetric cell exhibited a charge/discharge life of more than 2700 h at 1 mA cm⁻² and 1 mAh cm⁻² with a voltage hysteresis of about 37 mV. In brief, the superior cycling stability and reversibility of the Zn/MXene–40%CNF/Zn symmetric cell were attributed to the combined multifunctional effect of the MXene–CNF hydrogel interfacial layer, which provided strong stress confinement, low Zn²⁺ migration energy barriers, and abundant zincophilic sites. That is, this work provides a method to explore mechanically robust 1D/2D cellulose composite hydrogel interfacial layers to stabilize zinc anodes of zinc-ion batteries.

Compared to lithium-ion batteries, zinc–iodine batteries have emerged as ideal candidates for next-generation large-scale energy storage systems because of their environmental sustainability, inherent safety, and potential cost-effectiveness.^171–173 However, their application is limited by slow Zn²⁺ transfer kinetics, relatively low mass loading of iodine cathodes, and severe polyiodide shuttling. In this regard, Zheng and coworkers proposed a design strategy for Zn–I₂ batteries, by integrating a thick iodine cathode with a bacterial cellulose hydrogel electrolyte to form a continuous 3D ion transport network.⁷⁰ Notably, the polar BC fibers can form an interconnected network that provides abundant ion channels for inward transport of Zn²⁺ while limiting the solubilization of iodine species. The commercial BC hydrogels were immersed in 4 wt% NaOH at room temperature and high temperature treatment was carried out to ensure purification. The purified BC was freeze-dried and immersed in 2 M ZnSO₄ aqueous solution to prepare the bacterial cellulose hydrogel electrolyte. Notably, the shorter interaction distance and higher binding energy between iodine species (e.g., I₂, I₃⁻, and I⁻) and BC indicated that the adsorption of these iodine species on BC is thermodynamically favored, a property crucial for anchoring soluble iodine intermediates and preventing their migration during storage and recycling. Therefore, the BC hydrogel electrolyte significantly suppressed the growth of zinc dendrites in the assembled zinc–iodine batteries compared with the GF separator, and a uniform zinc surface was observed. In particular, the Zn/Zn symmetric batteries with BC hydrogel electrolytes sustained 480 h of stable cycling at 60% depth of discharge, while batteries with the GF separator failed after only 50 h. Moreover, the coulombic efficiency of a Zn/Cu asymmetric battery using BC hydrogel electrolyte also exceeded 99.6% at a current of 2–5 mA cm⁻², which also indicates the high reversibility of the Zn anode. Furthermore, a quasi-solid pouch cell at the 80 mAh-level was also able to cycle stably for more than 280 cycles at an iodine loading of 10 mg cm⁻². Indeed, this BC electrolyte design idea overcomes the trade-off between fast Zn²⁺ transport and high iodine loading, offering a highly promising strategy for fabricating high-energy-density quasi-solid-state zinc–iodine batteries.

5.3 Rechargeable sodium batteries

Compared to lithium and zinc metals, sodium metal is particularly abundant, exceeding lithium by 2–3 orders of magnitude. In particular, rechargeable sodium batteries, with a high theoretical specific capacity of 1165 mAh g⁻¹ and a low redox potential of −2.714 V, have attracted great attention from the academic community.^174,175 However, uneven distribution of concentration polarization and local current density lead to sodium anode volume expansion and dendritic crystal growth, which ultimately deteriorates the stability of the interface and causes short circuits.¹⁷⁶ Additionally, the Na anode readily reacts with organic electrolytes to form a loose solid electrolyte interphase (SEI) layer rich in organic components, resulting in a large interfacial impedance. Furthermore, the unstable SEI readily breaks down, exposing fresh Na metal, which then consumes more of the electrolyte until it is completely depleted.¹⁷⁷ That is, the above issues have a significant impact on the safety, performance, and lifetime of Na batteries and are key challenges that must be addressed before Na battery technology can move towards practical applications.

As we mentioned above, PEMs provide channels for ion transport as one of the key components of batteries and have a profound effect on the plating/stripping behavior at the sodium anode interface.¹⁷⁸ Indeed, the conventional commercial nonpolar polyolefin PEMs have poor compatibility with polar carbonate electrolytes, resulting in slow kinetic processes that lead to low ionic transport rates and inhomogeneous sodium deposition. Therefore, to modulate the transport behavior of Na⁺ and reduce the interfacial impedance of the Na anode, Xu and coworkers proposed an interfacial engineering strategy to construct a NaF-rich SEI for sodium-ion batteries using carboxyl-functionalized cellulose PEMs (Fig. 17c).¹⁷⁹ Notably, density functional theory calculation was first used to guide the selection of cellulose modification pathways. Moreover, the average electrostatic potential (ESP) value of the oxidized cellulose structural unit is −0.54 kcal mol⁻¹, which is much lower than the raw cellulose (−0.35 kcal mol⁻¹), aminated cellulose (1.22 kcal mol⁻¹), and sulfonated cellulose (−0.39 kcal mol⁻¹). Moreover, the oxidized cellulose structural units also showed the highest dipole moment, indicating that the oxidative modification gave cellulose a high atomic electronegativity difference. Therefore, the environmentally friendly and efficient 2,2,6,6-tetramethylpiperidinyloxy (TEMPO) oxidation method was subsequently employed for the modification of cellulose. Furthermore, oxidized cellulose (OC)-mediated PEMs were prepared by a wet process using oxidized cellulose as the raw material and aramid nanofibers (ANF) as fillers. Meanwhile, raw cellulose (RC)-based PEMs were prepared from unmodified raw cellulose as the control sample using the same preparation method. In particular, the symmetric cell showed an overpotential of 72 mV at 1400 h at 0.25 mA cm⁻², which was significantly better than the GF separator with uncontrolled Na dendrite-induced short circuits after 1017 h. Obviously, the Na/RC/Na symmetric cell showed voltage fluctuations after 1000 h, indicating its poor interfacial stability. In this process, the COOH group in OC PEMs promotes the cleavage of the P–F bond, thereby forming a NaF-rich SEI. In other words, the SEI with high NaF and low oligomers is smooth and strong, which can significantly reduce the interfacial impedance of the Na anode and prolong the lifetime of Na metal batteries.

Additionally, the diffusion of polysulfides and the low kinetics of the conversion reaction for sodium–sulfur batteries are the main challenges for their application. Specifically, similar to Li–S batteries, sulfur and sodium sulfide (Na₂S) have significant volume expansion during cycling, poor electrical conductivity, and the infamous “shuttle effect” caused by the diffusion of soluble sodium polysulfide (NaPS).^180–182 Consequently, chemical anchoring and/or physical confinement strategies are generally employed in sodium-sulfur batteries to improve sulfur utilization and inhibit polysulfide shuttling. In this regard, Peng and coworkers prepared a composite PEM to inhibit the sulfide shuttle effect in sodium–sulfur batteries by coating a 3D cellulose nanofiber-derived carbon aerogel onto a glass fiber separator (NSCA@GF).¹⁸³ Notably, the cellulose nanofiber-derived carbon aerogel separators with a hierarchical porous structure, good electronic conductivity, a 3D interconnection network, strong polysulfide anchoring capacity, and fast polysulfide conversion reaction kinetics can be used as a barrier layer and an enlarged current collector to improve sulfur utilization. The NSCA@GF PEMs were made by applying a thin layer of the NSCA membrane on top of the GF separator through a traditional papermaking process. A smooth surface and dense microstructure could be observed after GF was functionalized by NSCA, and the voids were completely covered by tightly arranged graphene sheets, thus forming a stable physical barrier to impede the penetration of NaPSs. In brief, NSCA functions mainly through the following aspects: (i) the NSCA layer has strong chemical absorption and physical shielding ability and can be used as a barrier layer for anchoring NaPS and (ii) the NSCA layer can act as a second current collector, improving the utilization of the active material due to its layered porous structure and good electron conductivity. Therefore, the battery with NSCA@GF had an initial discharge capacity of up to 952.2 mAh g⁻¹ at 0.5C and stabilized at 564.7 mAh g⁻¹ after 500 cycles, with a decay rate of only 0.08% per cycle and a coulombic efficiency of over 99%. In contrast, the initial discharge capacity of the battery with GF was 790.3 mAh g⁻¹ and only 100.3 mAh g⁻¹ remained after 500 cycles, with a capacity decay rate as high as 1.7% per cycle. This work demonstrated the application of cellulose nanofiber-derived carbon aerogels in sodium–sulfur batteries and also provided new ideas for the design of novel cellulose-mediated PEMs.

5.4 Rechargeable aluminum batteries

In addition to the rechargeable metal batteries discussed above, cellulose-mediated PEMs can also show unexpected effects and performance in rechargeable aluminum batteries.^184,185 As we know, aluminum (Al) is a resource-rich metal characterized by low cost, high stability, and mature mining/processing technology. In particular, it can be used as an anode to provide an ultra-high weight capacity (2980 mAh g⁻¹) second only to lithium metal (3861 mAh g⁻¹), with a volumetric capacity (8046 mAh g⁻¹) about four times that of lithium.¹⁸⁶ In other words, rechargeable aluminum batteries with cost-effectiveness, environmental sustainability, and high theoretical specific capacity have become a competitive alternative to lithium-ion batteries. Similarly, rechargeable aluminum batteries are also plagued by rapid dendrite growth and unstable physicochemical properties of the PEMs triggered by corrosive ionic liquid electrolytes (ILs). In the case of cellulose materials, the hydrogen bonding of cellulose is broken upon exposure to IL solutions, which can easily lead to membrane and fiber dissolution. Therefore, the solubilization of cellulose must be addressed to allow its stable application in rechargeable metal batteries.

To address the dissolution of cellulose in ionic liquids for its stable application in rechargeable aluminum-ion batteries, Xiong and coworkers prepared commercial cellulose paper@TiO₂ (CCP@TiO₂) with a core–shell structure by introducing TiO₂ nanolayers on commercial cellulose paper to solve the challenge of cellulose dissolution in ILs and thus improve its stability (Fig. 17d).¹⁸⁷ Notably, this intact and tightly shell-structured TiO₂ can act as a physical shield to prevent the nuclear structured CCP from coming into contact with the IL and protect the H-bonds between the cellulose inside the CCP, leading to an extremely stable PEM in the IL while maintaining an intact fibrous and pore structure, which can contribute to uniform and stable ion channels and anion fluxes. Moreover, the shell-like structure of TiO₂ can produce a polarization effect at the electrolyte/separator interface, which promotes a uniform distribution of electric field across the electrolyte/anode/separator surface, which in turn enhances the ionic conductivity and the amount of anion transport and ultimately achieves a dense, dendrite-free planar Al deposition. Benefiting from the superb stability of CCP@TiO₂ in ILs, CCP@TiO₂ exhibited prominent stability in rechargeable Al batteries. Compared to GF (459 h) and pure CCP (312 h), Al//Al symmetric batteries assembled based on CCP@TiO₂ exhibited significantly longer cycle life (over 800 h) and more stable flat voltage plateau, demonstrating superior long-term Al plating/stripping cycling performance at 0.5 mA cm⁻² and 1 mAh cm⁻². Meanwhile, CCP@TiO₂ exhibited higher Al plating/stripping current densities in the Al//Mo half batteries compared to GF, suggesting that CCP@TiO₂ provided better Al plating/stripping reaction kinetics.

Al–air batteries are typically made of hard and heavy materials and also require a large amount of static or cycled electrolyte solution to ensure their stable operation.¹⁸⁸ Similarly, there are several key areas requiring focused attention during the discharge process of Al–air batteries: (i) side reactions at the electrode–electrolyte interface consume the Al anode and (ii) oxygen levels in ambient air are constantly decreasing due to concentration gradients. Unlike conventional rigid batteries, flexible batteries can be used in complex environments where the shape of the battery is less demanding. Especially considering the tremendous advantages of Al in terms of energy density, cost, abundance, and environmental friendliness, flexible Al–air batteries are highly promising and competitive in powering a wide range of disposable devices.¹⁸⁹ In this regard, Wang and coworkers fabricated a novel flexible Al–air battery that was prepared entirely on cellulose paper.¹⁹⁰ Specifically, the Al foil anodes were embedded within the paper during the papermaking process, while the permeable cathodes were deposited directly onto the paper surface through redox reactions with ink. Notably, the Al foil was nicely sandwiched between two layers of cellulose paper, and the ORR ink did not penetrate the paper layers or come into contact with the Al anode, thus eliminating the issue of a shorted cell. Moreover, a thin gap was observed between the paper envelope and Al foil, demonstrating that there is no seamless contact between them. As the paper-based PEMs absorb the electrolyte, the cellulose fibers expand slightly to clamp the Al foil tightly, thus ensuring proper battery operation. Furthermore, the electrode area of the assembled Al–air battery was enlarged by 5 times using 4 M NaCl as the electrolyte, the current output could be increased from 40 mA to 200 mA, and the peak power was increased from 8.0 mW to 37.4 mW. Meanwhile, the Al–air battery with an electrode area of 10 cm × 1 cm, using cellulose paper as a separator, was able to power a small fan even when rolled into a spiral, further demonstrating the battery's excellent flexibility. Furthermore, in the case of Al–air batteries, discharge life and Al utilization efficiency are significantly reduced due to unavoidable Al corrosion. For example, a battery featuring 3.5 mg of the Al anode and a NaCl/NaOH electrolyte can discharge stably at current densities ranging from 1 to 20 mA cm² until Al is depleted, causing a sudden voltage drop. In comparison, alkaline electrolytes were more suitable for short-term operation with higher voltage requirements for cellulose paper-based PEM-assembled Al–air batteries, while brine electrolytes were more suitable for long-term tasks with lower voltage requirements.

5.5 Supercapacitors

Flexible supercapacitors are a promising avenue for unlocking the potential of flexible electronics, with high power density, excellent cycling stability, and long operating life and are considered an ideal power source for next-generation wearable devices.^191–193 Notably, to construct high-performance flexible supercapacitors for practical applications, extensive efforts have been made to optimize the specific capacitance of electrodes, the gel electrolyte, and the ionic conductivity of the PEMs in the electrolyte.^194–196 Despite several impressive advances, the weak bond between electrodes and gel electrolyte/separator makes it difficult to avoid mechanical mismatch under a variety of flexibility conditions. Therefore, the integration of various components into high-performance flexible supercapacitors remains a major challenge.

To realize the mechanical and electrochemical stability of flexible supercapacitors under different flexibility conditions, inspired by the multilayered continuous structure of natural biological tissues, Yu and coworkers developed an integrated supercapacitor (BA-SC) based on bacterial cellulose by sequentially constructing a continuous electrode–separator–electrode structure in an uninterrupted biosynthetic process (Fig. 18a).¹⁹⁷ Specifically, the bottom electrode layer, the middle separator layer, and the top electrode layer were constructed in following stages: (i) the liquid medium is converted to an aerosol during the construction of the separation layer and is then fed into the process of synthesizing pure BC; (ii) activated carbon and carbon black particles were added to disperse the aerosol during the construction of the electrode layer, so that the deposition of the activated carbon and carbon black particles and the growth of the BC were spatially and temporally coupled with each other. Therefore, activated carbon and carbon black particles are entangled in situ in a three-dimensional network of BC to form a uniform electrode layer. In particular, the biosynthesis of BC was not interrupted throughout the aerosol-assisted biosynthesis, ensuring that the three-dimensional network of BC continued to grow during the phase change. Therefore, the obtained BA-SC cross-sectional image showed a clearly visible three-layer structure (electrode–separator–electrode) with no delamination between neighboring layers. Precisely because of this tight and consistent structure, the obtained BA-SC has superior resistance to shear and bending compared to supercapacitors without a three-layer continuous structure, while maintaining electrochemical properties. Moreover, this three-layer continuous structure based on bio-inspiration can be easily assembled into symmetric supercapacitors, thus exploring its competitiveness in capacitive energy storage. The cyclic voltammetry curves of BA-SC exhibited an almost mirror-image symmetric quasi-rectangular shape at different scan rates from 2 to 200 mV s⁻¹, which indicated near-ideal capacitive behavior and efficient electrolyte ion transport. Furthermore, the potential window was from 0 V to 1.0 V at current densities from 0.5 to 10 A g⁻¹, and the electrostatic charge/discharge curves also showed typical symmetrical triangular profiles, which were consistent with the analysis of CV curves. Additionally, BA-SC exhibited excellent mechanical properties compared to conventional supercapacitors assembled in layers due to the three-dimensional network formed by BC throughout the materials. Noteworthily, this fast ion diffusion and stable mechanical structure contribute to high specific capacitance and long-term stable electrochemical performance at high loading mass, respectively.

Fig. 18 Application of cellulose-mediated PEMs in novel energy equipment. (a) Schematic diagram of the fabrication of supercapacitors using bacterial cellulose (BA-SC) and the cycling stability along with electrostatic charge–discharge curves (reproduced with permission from ref. 197. Copyright 2024, Wiley-VCH). (b) Schematic diagram of the preparation process of a quaternized CNC (QCNC) and the illustration of hydrophilic channels in composite AEMs (reproduced with permission from ref. 207. Copyright 2018, American Chemical Society). (c) Schematic diagram of PSCs passivated with cellulose derivatives and power conversion efficiencies (reproduced with permission from ref. 211. Copyright 2024, Springer Nature). (d) Schematic diagram of MoS₂ assembled with cellulose nanofibers into a membrane and the power output (reproduced with permission from ref. 214. Copyright 2021, American Chemical Society).

The characteristics of supercapacitor PEMs virtually determine the wearable characteristics of the capacitor, including its compactness, reliability, and freedom from leakage. In general, the PEMs used in these flexible supercapacitors are typically a hydrogel electrolyte consisting of a polymer physical framework injected with an electrolyzed salt/acid/base solution.^41,198,199 Indeed, most hydrogel electrolytes are only used as bendable supports with poor electrolyte ion storage capacity and slow ion transport kinetics compared to liquid electrolytes. In other words, it remains a challenge to prepare device-grade hydrogel electrolytes that are both flexible enough to accommodate a wide range of deformations and tough enough, meaning that they are not vulnerable to weak mechanical properties. In this regard, Guo and coworkers designed an all-solid-state supercapacitor utilizing graphene-wrapped polyester fibers as flexible electrodes and polyacrylamide-reinforced bacterial cellulose nanofibers as a hydrogel electrolyte (BC/PAM).²⁰⁰ In brief, the physical interaction between the amino group of PAM and the hydroxyl group of BC contributes to the formation of a three-dimensional hydrogel. In this process, the critical step in BC/PAM is the introduction of BC nanofibers into an aqueous solution of N,N′-methylenebisacrylamide (MBAA), ammonium persulfate, and acrylamide before the polymerization process. The BC nanofibers were embedded into the PAM framework by free-radical polymerization using MBAA as a cross-linking agent during the thermal processing. The well-embedded BC can transfer mechanical stress from PAM chains to nanofibers, resulting in a reinforcing effect. Therefore, the BC/PAM hydrogels with good strength exhibited a strain stretch of 1300% without significant cracking or fracture. This hybrid hydrogel could be deformed arbitrarily and had a high elongation even in the knotted state. In particular, this structural integrity of BC/PAM conferred a low interfacial resistance, which facilitated the rapid transport of ions across the interface during continuous mechanical deformation. Thus, the supercapacitor assembled with BC/PAM exhibited a wide voltage range and current density at different scan rates. Meanwhile, the supercapacitor obtained a high areal capacitance of 564 mF cm⁻² at a current density of 1 mA cm⁻², and it maintained 437.5 mF cm⁻² at the high current density of 50 mA cm⁻². Additionally, the integrity of the capacitor was well maintained thanks to the good individual components, which greatly enhanced its mechanical stability in the case of bending deformation without affecting the overall electrochemical performance.

5.6 Fuel cells

Among the various advanced energy storage and conversion devices being explored to address energy and environmental issues, proton exchange membrane fuel cells have been widely recognized as a highly promising option because of their excellent performance in terms of high energy density, mild operating conditions, and low pollutant emissions.^201,202 Despite extensive research efforts, designing PEMs with high proton conductivity and stability for fuel cells remains a challenge. The commonly utilized or investigated PEMs, such as Nafion and its replacement polymers, inorganic/polymer composites, and porous inorganic/carbon materials, typically exhibit a significant reduction in proton conductivity under relatively low humidity conditions.^203–205 Meanwhile, developing biodegradable PEMs for fuel cells has become a significant challenge as the environmental sustainability of separator materials gains increasing attention.

Cellulose nanocrystals (CNC), with the advantages of low cost, biodegradability, and easy chemical modification, are a very promising alternative material for Nafion membranes and have received tremendous attention. Nevertheless, PEMs prepared directly from CNC have low proton conductivity and low mechanical properties and cannot meet the requirements of fuel cell applications. In this regard, to construct CNC-mediated PEMs with excellent proton conductivity and stability, Zhao and coworkers grafted cysteine (Cys), taurine (TAU), and cytidine monophosphate (CMP) onto CNC to prepare a novel CNC-mediated PEM for methanol fuel cells by vacuum-assisted self-assembly.²⁰⁶ Specifically, the grafting method consists of two processes: (i) NaIO₄ acts as an oxidant to oxidize CNC to 2,3-dialdhyde cellulose (DAC) and (ii) acetic acid is used as a catalyst to allow the nucleophilic primary aldehyde groups on DAC to undergo a Schiff base reaction with modified chemicals. Moreover, the aldehyde group on DAC can also cross-link with the hydroxyl group on the CNC to form a hemiacetal, which can enhance the tensile properties of CNC-mediated PEMs. Notably, the power output of the Nafion 117 membrane was 34.95 mW cm⁻² at 80 °C, while the pure CNC membrane had no power output due to the lack of ion-conducting groups. In particular, the power density of the modified CNC composite membrane increased with the content of the modified materials, because the more the phosphate groups or sulfonic acid groups, the more the proton transport sites. For example, the maximum power density of the TAU3-DAC membrane was 34.05 mW cm⁻², which was 97.4% of the Nafion 117 membrane.

Generally, one way to improve the ion conductivity of proton exchange membrane is to increase the fixed cation concentration within the membrane. In this case, the mechanical stability of PEMs is compromised due to excessive water absorption. Therefore, cross-linking is an effective method for preparing PEMs with high ionic conductivity and robustness. However, ion conduction paths may be blocked by the cross-linked network; thus, balancing cross-linking degree and conductivity is particularly important. Therefore, to improve the ionic conductivity and stability of cellulose-mediated PEMs for fuel cells, Zhang and coworkers utilized quaternary ammonium groups to surface-modify cellulose nanocrystals (CNC) and incorporated them into quaternized poly(phenylene oxide) (QPPO) to make PEMs with high ionic conductivity and stability for fuel cells (Fig. 18b).²⁰⁷ Specifically, the original CNC was surface esterified with negatively charged sulfate groups by hydrolysis with sulfuric acid. Furthermore, the hydrolysis of the alkoxy of amino silane resulted in a dehydration condensation reaction with the hydroxyl group of CNC, which resulted in the surface functionalization to an amino group. Finally, the amino group was converted to a quaternary ammonium group by iodomethane, resulting in a QCNC with a surface cationic group. Moreover, the different surface morphology of the QCNC compared to the CNC means that silane has successfully formed a coating on the CNC surface. Notably, the QCNC is highly hydrophilic due to the abundance of hydroxyl and cationic groups on its surface, which can attract surrounding water molecules and thus simultaneously form hydrophilic channels in the PEMs, promoting ionic conduction and improving the conductivity of PEMs. Therefore, the QCNC composite membrane doped with 2wt% has higher ionic conductivity stability and durability compared to the pristine QPPO membrane. Furthermore, the membrane electrode assembly (MEA) based on neat QPPO exhibited a peak power density of 270 mW cm⁻² at 0.5 V and 0.551 A cm⁻². In comparison, the power density of the 2 wt% QCNC/QPPO-based MEA increased to 392 mW cm⁻² at 0.85 V and 0.45 A cm⁻², which was 145% of QPPO-based MEA. That is, the modified QCNC can effectively enhance the fuel cell performance of the PEMs.

5.7 Solar cells

Perovskite materials for solar cells are gaining increasing attention due to their excellent absorption coefficient, high carrier mobility, and low exciton binding energy.^150,208,209 Currently, the highest certified power conversion efficiency (PCE) value for solar cells reached 26.1%, which is still below the theoretical limit of 33.7% based on the Shockley–Queisser limit.²¹⁰ Notably, defects tend to form habitually during the fabrication of perovskite materials, such as interface defects, bulk defects, and grain boundary defects. Defect formation not only causes photogenerated carriers to release energy through non-radiative recombination but also hinders the accumulation of electrons and holes at the electrodes, thereby suppressing improvements in the device's open-circuit voltage and fill factor. Cellulose chains exhibited excellent grain boundary enrichment capabilities, enabling defect passivation and promoting grain growth through hydrogen bonding interactions between perovskite structures and hydroxyl groups. Therefore, to further investigate the interaction relationship and mechanism between the cellulose structure and the perovskite, Zhang and coworkers used natural cellulose as a raw material to design a cellulose derivative for perovskite crystallization engineering (Fig. 18c).²¹¹ In this process, the cationic cellulose derivative C-Im-CN containing the imidazolium cyanide cation (Im-CN) and chloride anion significantly promotes the crystallization process, grain growth, and orientation of perovskites. Specifically, the ionic cellulose derivative C-Im-CN can provide multiple interaction sites, including imidazolium cations, cyano groups, hydroxyl groups, carbonyl groups, and chloride anions, which can strongly interact with perovskites through electrostatic forces, coordination interactions, and hydrogen bonding. Therefore, perovskite solar cells based on C-Im-CN mediation demonstrate a high power conversion efficiency of up to 24.71%. In particular, the unencapsulated perovskite solar cells maintained over 91.3% of their initial efficiency after continuous operation for 3000 h in conventional ambient air, demonstrating excellent stability under high humidity conditions.

5.8 Reverse electrodialysis

Reverse electrodialysis is considered a highly promising technology capable of directly converting the electrochemical potential difference of a salinity gradient into electrical energy, but its application is primarily constrained by the relatively low performance of its core membrane components.^212,213 In this regard, the ultrafast permeation and highly selective transport properties of 2D nanofluids hold immense potential for overcoming the limitations of conventional membrane technologies, opening up entirely new avenues for harvesting permeation energy. Recently, the potential of various two-dimensional materials such as MXenes, black phosphorus, boron nitride, and graphene has been successively explored, leading to the development of several advanced membrane design concepts. Nevertheless, the maximum power density of half-cell systems remains limited to around 4 W m⁻², below the commercialization benchmark (more than 5 W m⁻²). Therefore, to prepare ion-exchange electrolyte materials with high energy density, Zhang and coworkers assembled MoS₂ nanosheets with cellulose nanofibers to form a high-strength and robust 2D MoS₂/CNF composite PEM (Fig. 18d).²¹⁴ Thanks to the introduction of CNF as a reinforcing agent, the MoS₂/CNF composite PEM can combine the advantages of high ionic conductivity with superior mechanical properties, offering a new research direction for permeation energy conversion technology. In brief, the introduction of CNFs enhances the ion selectivity and physical steric hindrance effect of composite materials. Notably, the MoS₂/CNF composite PEM can achieve an output power density of 5.2 W m⁻² when artificial river water mixes with seawater. Additionally, the MoS₂/CNF composite PEMs can work stably in natural water environments and exhibit an increased power output of approximately 6.7 W m⁻², and their permeate energy conversion performance in both natural and artificial water sources not only surpasses commercial ion-exchange membranes but also surpasses the current state-of-the-art macroscale two-dimensional nanofluidic membranes.

Following a similar design approach, Chen and coworkers prepared a novel cellulose-mediated electrolyte material by integrating 1D bacterial nanofibers with 2D graphene oxide/layered double hydroxide nanosheets (BC/GO, BC/LDH), achieving significantly enhanced power density.²¹⁵ Specifically, BC/GO and BC/LDH with outstanding stability and ion selectivity were prepared by vacuum-assisted filtration after dispersing BC and GO, BC and LDH in varying proportions. In particular, the prepared BC/GO and BC/LDH can form superimposed electrochemical potentials through complementary diffusion of oppositely charged ions, achieving an output power density as high as 0.7 W m⁻² when using artificial seawater and river water. By connecting twenty units, the series-connected reverse electrodialysis device can achieve an output voltage of up to 2.2 V. Briefly, this work provided a viable solution for balancing ion selectivity and permeability by employing a strategy of mixed-dimensional assembly between 1D BC and 2D nanosheets, offering a universal technical pathway to address the challenges faced by salinity gradient power generation technology.

Polymer electrolyte materials constructed from cellulose exhibit tremendous application potential in numerous energy storage and conversion devices due to their advantages of being renewable, biodegradable, mechanically robust, and easily functionalized. However, a series of critical, common, and specific scientific and engineering challenges must still be addressed to bridge the gap between outstanding laboratory performance and practical large-scale applications: (i) long-term cycling and structural stability: the long-term chemical, structural, and electrochemical stability of cellulose-based PEMs must be tested under realistic conditions such as charging/discharging cycles, temperature variation, and vibration, rather than being confined solely to the laboratory stage. Meanwhile, achieving continuous, defect-free, and thin-film production of cellulose-mediated PEMs is essential, while ensuring seamless compatibility with existing battery industry manufacturing processes such as electrode coating and roll-to-roll coating. Furthermore, consistency and reproducibility in both cellulose raw material and final product performance must be guaranteed. (ii) Customized designs for different applications: in applications such as rechargeable batteries, cellulose-mediated PEMs require the transport of ions with diverse properties, necessitating precise design of the pore structure and physicochemical properties. A precise regulation of cellulose-mediated PEM performance requires an integrated approach spanning molecular structure design of cellulose, control of micro-scale morphology and macro-scale device integration.

6. Environmental sustainability and industrial scalability

Growing environmental concerns require the substitution of petroleum with sustainable and renewable resources to fabricate biodegradable and carbon-neutral products. In this regard, the urgency of phasing out fossil carbon and limiting anthropogenic greenhouse gas emissions from fossil fuels is undeniable. Developing energy-related materials that are both environmentally friendly and electrochemically competitive, while balancing the conflicting aspects of renewability, material abundance, electrochemical performance competitiveness, and low carbon footprint, represents one of the most challenging tasks in the field of energy storage.^34,216 Cellulose-mediated PEMs can play a vital role in the transition toward renewable carbon-based chemicals and derived materials (Fig. 19a).²¹⁷ In brief, cellulose can enhance carbon circularity by substituting fossil fuel feedstocks and integrating biomaterials into technological cycles, thereby preventing fossil carbon from entering the technosphere, biosphere, and atmosphere (de-fossilization strategy) (Fig. 19b). Moreover, cellulose-mediated PEMs can function as carbon sinks during application, with their bio-based carbon accumulating in soil after biodegradation via β-1,4-glycosidic bond hydrolysis. Cellulose can be obtained year-round from abundant renewable lignocellulosic crops and non-food crops and can also be extracted from second-generation and third-generation feedstocks, thereby enhancing circularity and reducing waste. Specifically, the carbon footprint of cellulose fibers (e.g., kenaf, hemp, jute, and flax) and their derivatives (e.g., carboxymethyl cellulose) ranges from approximately 0.3 to 4.7 CO₂ equivalents per kilogram (eqkg⁻¹), significantly lower than that of carbon fibers derived from polyacrylonitrile (25–43 eqkg⁻¹). In terms of energy consumption during material preparation, cellulose-mediated PEMs obtained through molecular engineering demonstrate certain energy-saving advantages: the energy consumption for dehydrating micro/nanocellulose or cellulose nanocrystals via ultrasonic or electro-osmotic methods ranges from 0.6 to 2.4 MJ kg⁻¹, and certain cellulose-mediated PEMs can be formed simply by mixing and drying at room temperature.

Fig. 19 The circularity and environmental impact of cellulose-mediated PEMs. (a) Comparison of carbon distribution in various forms in 2020 and 2025 (the yellow area represents the carbon content in biomass) (reproduced with permission from ref. 34. Copyright 2025, Springer Nature). (b) The cyclic properties of cellulose-mediated PEMs throughout their entire lifecycle. (c) Life cycle assessment of cellulose-mediated PEMs. (reproduced with permission from ref. 216. Copyright 2025, Wiley-VCH).

Additionally, life cycle assessment (LCA) can be employed to measure the cumulative environmental impact of the cellulose manufacturing process, revealing its relative environmental footprint compared to precursor raw materials and providing a quantitative reference basis for evaluating the environmental attributes of different materials with similar application scenarios (Fig. 19c). Briefly, the LCA of cellulose can be investigated from ozone depletion, water depletion, fossil fuel depletion, human toxicity, climate change, cumulative energy demand, and terrestrial acidification. By analyzing the multifaceted impacts of cellulose during production, potential environmental hotspots that may have been overlooked can be identified. Such information contributes to identifying environmental hotspots and optimizing materials to reduce resource, water, and energy consumption, thereby minimizing carbon emissions and other environmental impacts such as acidification, eutrophication, and particulate matter formation. However, it is currently impossible to accurately assess the comprehensive environmental impacts of cellulose products across different areas such as human health, air emissions, and waste stream discharges, as only laboratory data are currently available. Moreover, data unavailability represents another limitation of laboratory-scale processes, as these procedures fail to account for the LCA of cellulose-based materials. That is, laboratory-scale synthesis and preparation must be closely integrated with industrial-scale production to ensure consistency in material selection and energy recovery throughout the process. In the case of the chemical-mechanical preparation route, the majority of cellulose's environmental impact during preparation depends on the chosen chemical modification and mechanical processing pathways. For example, the mechanical processing involved in ultrasonic treatment may mask the effects of chemical modification in certain cellulose processing methods. For derivatives of cellulose, the optimal implementation method involves homogenization following TEMPO oxidation, as the environmental impact of TEMPO oxidation is lower than that of carboxymethylation reactions.

It is worth noting that in the process of advancing toward practical industrial applications, scaling up production and controlling costs are critical issues that cannot be avoided. Derived from wood, plant straw, and cotton linters, cellulose stands as nature's most abundant biopolymer with renewable, biodegradable, and low-cost properties, offering significant advantages over traditional petroleum-based polymer feedstocks or specialty engineering plastics. For example, the synthesis of poly(acrylonitrile) (PAN) and poly(ethylene oxide) (PEO) relies on petrochemical feedstocks, whose prices are highly susceptible to fluctuations in international crude oil prices. In contrast, cellulose, as a natural polymer, primarily derives from renewable plant resources, resulting in relatively stable production costs. From a cost-benefit analysis of transitioning from cellulose to PEMs, the low cost of cellulose feedstock itself represents its core advantage, but processing costs currently constitute the primary bottleneck in achieving low-cost production.²¹⁸ From a manufacturing perspective, compared to other preparation methods, the solution recasting technique offers convenient control over the thickness and shape of cellulose-mediated electrolyte materials while requiring relatively less sophisticated equipment, making it a more promising option for practical applications. Moreover, the organic reagents and acid-base reagents involved in designing cellulose-mediated PEMs may significantly impact the durability of industrial production equipment. Therefore, developing greener alternative solvents, optimizing process parameters to reduce energy consumption, and exploring compatibility with existing polymer processing technologies are critical steps toward achieving economically viable large-scale production. To further advance the industrial applications of cellulose-mediated PEMs, efforts should focus on raw material sourcing, preparation processes, large-scale production strategies, and unified evaluation metrics, with comprehensive technical-economic analysis and life cycle assessment conducted.

7. Conclusion and perspectives

Despite the great progress and development of cellulose towards novel PEMs through strategies such as molecular engineering, there are still significant research challenges to bridge the gap between existing capabilities and practical applications for performance adaptability and sustainable systems (Fig. 20).

Fig. 20 A summary of challenges/solutions, key requirements, characterization techniques, and future development of cellulose-mediated PEMs.

(i) Pre-treatment of cellulose: influenced by intramolecular and intermolecular non-covalent forces, cellulose fibers can be divided into crystalline and non-crystalline regions, making them difficult to dissolve in conventional solvents. The current mainstream cellulose solvents are NaOH/urea solutions or ionic liquids, which suffer from unwanted residues and high costs, respectively. For industrial applications, the current cellulose dissolution technologies are mainly NaOH/CS₂ and N-methylmorpholine-N-oxide systems (NMMO), which also suffer from the defects of the complicated process, unfriendly environment, and high price, respectively. Therefore, the development and design of cellulose dissolution systems using a simple process, low price, high efficiency, and stability is a key step to promote cellulose-mediated PEMs towards industrial applications and a breakthrough in the development of cellulose-mediated functional materials. Moreover, the modification of cellulose by molecular engineering using reactions such as etherification, esterification, oxidation, etc. is another feasible way to solve the insolubility problem and also give it unique properties (e.g., superior ionic conductivity, mechanical properties, porosity, flame retardant, self-healing, and solubility resistance).

(ii) Iteration of preparation strategies: the current preparation strategies for cellulose-mediated PEMs mainly depend on solution casting/coating, vacuum-filtration, electrostatic spinning, and phase inversion. Indeed, different preparation strategies need to be strictly matched to the type and physicochemical characteristics of cellulose. In view of the higher demands on the performance of cellulose-mediated PEMs for new applications, the preparation strategies for cellulose-mediated PEMs need to be iterated. For example, the combination of new strategies with the help of 3D printing, electrospray, in situ polymerization, photoinduced synthesis, and the pre-treatment technology of cellulose is collectively advancing the development and applications of cellulose-mediated PEMs. Especially for some special materials or properties that cannot be obtained by conventional methods, the exploration of new technologies will be a key step to unlock new cellulose-mediated PEMs in the future. Notably, these new technologies should be matched with industrialized equipment in order to address the impediments to the applications of cellulose-mediated PEMs.

(iii) Optimization of design concepts: multiscale rational material design for optimizing the structure–property–function relationship of cellulose-mediated PEMs required careful consideration of their intended use. Techniques that combine molecule-level molecular engineering with mesoscale and macroscale fabrication processes can produce cellulose-mediated PEMs with advanced functionality. In other words, the synergistic combination of multiscale modeling and experimental techniques can provide a possible way for molecular sequence engineering of cellulose-mediated PEMs for large-scale composite manufacturing. In this process, achieving precise control over the molecular structure of cellulose and ensuring consistency in cellulose-mediated PEMs is a key challenge. This requires meticulous control of the distribution of cellulose in the composite and careful management of the molecular interactions that determine the overall properties. A fine balance between interfacial bonding, compatibility, and molecular structure needs to be ensured in the design concept of cellulose-mediated PEMs.

(iv) Development of characterization techniques: to further understand the structure–property–function relationships of cellulose-mediated PEMs, advanced characterization techniques, as well as powerful modeling tools and theories, are needed at small and high-resolution time and length scales. For instance, in situ and real-time analysis of non-covalent interactions during the processing of cellulose-mediated PEMs can contribute to determining the dynamic assembly behavior and optimizing the process to obtain the desired molecular structure. Moreover, theoretical computational capabilities, including accurate toolkits and scaling up modeling frameworks, are other challenges that need to be overcome to predict function-specific outcomes. In particular, the multidimensional interactions, ionic conduction behavior, and interfacial interaction relationships within cellulose-mediated PEMs need to be quantified by accurate simulations to better reflect the structure–property relationships. Alternatively, machine learning and artificial intelligence may accelerate the optimization and discovery of cellulose-mediated PEMs for molecular engineering.

(v) Focus on long-term and sustainable development: despite demonstrating significant potential in terms of sustainability and mechanical properties, cellulose-mediated PEMs face two major bottlenecks in their transition to practical applications: insufficient ionic conductivity at room temperature and long-term stability issues in complex electrochemical environments. Future research should focus on the following specific directions: (i) through multidimensional molecular design, stable and highly flexible ion-conducting segments are grafted onto a rigid cellulose, aiming to decouple segment motion from ion transport and create a localized fast-ion environment. The formation of ordered nanostructures is explored through self-assembly by creating interpenetrating networks where cellulose serves as a rigid scaffold and highly conductive soft polymers form interpenetrating networks, thereby constructing distinct “ionic channels” or “permeation networks”. (ii) By introducing chemically sensitive bonds responsive to specific stimuli into the cellulose framework structure, novel cellulose-mediated PEMs are designed to maintain stability during operation while rapidly degrading under specific processing conditions, meanwhile advocating for the consideration of recycling and circularity from the initial stages of material design and researching efficient green solvent systems to achieve selective separation and recovery of cellulose matrices and other components.

Future advances in cellulose-mediated PEMs will require careful trade-offs between performance, environmental sustainability, and cost-effectiveness. Laboratory development of prototype materials is often guided by the principle of prioritizing performance at the expense of other important aspects, such as conducting a comprehensive life-cycle assessment of raw materials from procurement to end-of-life and minimizing waste of chemicals or water. Indeed, materials research and development efforts need to be interdisciplinary, incorporating insights from materials science, engineering, economics, and environmental science to create solutions that meet these diverse criteria. Despite the numerous challenges that remain, the sustainability and ease of modification of cellulose provide a broad stage for its utilization in PEM applications and highlight a fascinating and bright future.

Conflicts of interest

The authors declare no conflict of interest.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Supplementary information is available. Comparative data on cellulose-mediated PEMs and other electrolyte materials have been presented in the supporting information. See DOI: https://doi.org/10.1039/d5ee05398f.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (32501592 and 32471806), the Natural Science Foundation of Tianjin (25JCQNJC01060, 24JCZDJC00630, and 23JCZDJC00630), the China Postdoctoral Science Foundation (2023M740563), the Young Elite Scientist Sponsorship Program by Cast (No. YESS20230242), the Tianjin Enterprise Technology Commissioner Project (25YDTPJC00690), and the China Scholarship Council (No. 202408120105).

References

1. L. Yan, A. J. Huertas-Alonso, H. Liu, L. Dai, C. L. Si and M. H. Sipponen, Chem. Soc. Rev., 2025, 54, 6634–6651.
2. Q. Zhang, S. Dong, P. Shao, Y. Zhu, Z. Mu, D. Sheng, T. Zhang, X. Jiang, R. Shao, Z. Ren, J. Xie, X. Feng and B. Wang, Science, 2022, 378, 181–186.
3. Z. H. Wang, H. Hölzel and K. Moth-Poulsen, Chem. Soc. Rev., 2022, 51, 7313–7326.
4. Y. W. Xu, X. He, X. F. Liang, H. Huang, J. E. Wu, D. F. Ju, J. H. Guo, S. X. Dong, Z. Q. Hu and M. Liu, Energy Environ. Sci., 2025, 18, 8339–8351.
5. L. Y. Zhu, Y. Cao, T. Xu, H. B. Yang, L. Y. Wang, L. Dai, F. S. Pan, C. J. Chen and C. L. Si, Energy Environ. Sci., 2025, 18, 5675–5739.
6. W. Li, Y. Xu, G. H. Wang, T. Xu and C. L. Si, Adv. Energy Mater., 2024, 14, 2403593.
7. C. L. Song, L. M. Pan, L. Chen, Y. X. Jiang, H. J. Pan, H. Zhao, N. S. Chen, Z. Q. Yang, L. Yang, Q. Yan, X. D. Peng, X. H. Ma, Y. J. Li and T. S. Zhao, Energy Environ. Sci., 2025, 18, 8244–8255.
8. K. Zhao, X. X. Wang, W. F. Feng, A. Ponrouch, P. Poizot, M. Armand, Z. B. Zhou and H. Zhang, Energy Environ. Sci., 2025, 18, 7756–7791.
9. Y. X. Wang, T. Xu, K. Liu, M. Zhang, Q. S. Zhao, Q. D. Liang and C. L. Si, Ind. Crop. Prod., 2023, 204, 117378.
10. T. Xu, Q. Song, K. Liu, H. Y. Liu, J. J. Pan, W. Liu, L. Dai, M. Zhang, Y. X. Wang, C. L. Si, H. S. Du and K. Zhang, Nano-Micro Lett., 2023, 15, 98.
11. M. L. Wu, X. Zhang, Y. Zhao, C. P. Yang, S. S. Jing, Q. S. Wu, A. Brozena, J. T. Miller, N. J. Libretto, T. P. Wu, S. Bhattacharyya, M. N. Garaga, Y. G. Zhang, Y. Qi, S. G. Greenbaum, R. M. Briber, Y. S. Yan and L. B. Hu, Nat. Nanotechnol., 2022, 17, 629–636.
12. S. M. Hao, J. X. Zhu, S. He, L. Ma, W. Z. Liu, Y. Y. Zhang, X. X. Xie, X. Qin, X. L. Fan, H. Li, L. Q. Zhang and W. D. Zhou, Nat. Sustainable, 2024, 7, 661–671.
13. Q. Zhao, S. Stalin, C. Z. Zhao and L. A. Archer, Nat. Rev. Mater., 2020, 5, 229–252.
14. Q. Zhou, J. Ma, S. M. Dong, X. F. Li and G. L. Cui, Adv. Mater., 2019, 31, 1902029.
15. Z. H. Wang, Y. H. Lee, S. W. Kim, J. Y. Seo, S. Y. Lee and L. Nyholm, Adv. Mater., 2021, 33, 2000892.
16. W. Yang, W. Yang, J. M. Zeng, Y. L. Chen, Y. F. Huang, J. Liu, J. Y. Gan, T. Z. Li, H. Zhang, L. X. Zhong and X. W. Peng, Prog. Mater. Sci., 2024, 144, 101264.
17. C. J. Chen and L. B. Hu, Acc. Chem. Res., 2018, 51, 3154–3165.
18. X. C. Shi, Z. B. Wang, S. Y. Liu, Q. Q. Xia, Y. Z. Liu, W. S. Chen, H. P. Yu and K. Zhang, Nat. Sustainable, 2024, 7, 315–325.
19. T. Xu, K. Liu, N. Sheng, M. H. Zhang, W. Liu, H. Y. Liu, L. Dai, X. Y. Zhang, C. L. Si, H. S. Du and K. Zhang, Energy Storage Mater., 2022, 48, 244–262.
20. X. Y. Li, H. M. Wang, Y. C. Han, J. Li, M. C. Yang, Q. X. Hou and C. L. Si, Appl. Energy, 2025, 401, 126711.
21. T. Li, C. J. Chen, A. H. Brozena, J. Y. Zhu, L. X. Xu, C. Driemeier, J. Q. Dai, O. J. Rojas, A. Isogai, L. Wågberg and L. B. Hu, Nature, 2021, 590, 47–56.
22. H. L. Zhu, W. Luo, P. N. Ciesielski, Z. Q. Fang, J. Y. Zhu, G. Henriksson, M. E. Himmel and L. B. Hu, Chem. Rev., 2016, 116, 9305–9374.
23. H. B. Yang, L. Y. Zhu, Y. J. M. Zhou, T. Xu, C. Y. Zheng, Z. H. Yuan and C. L. Si, Chem. Eng. J., 2024, 498, 155333.
24. H. W. Ma, Z. Y. Cheng, X. B. Li, B. Li, Y. J. Fu and J. C. Jiang, J. Bioresour. Bioprod., 2023, 8, 15–32.
25. H. Zhou, Y. Guan, T. Xu, D. Tang, X. Yu, L. Dai, W. Chen and C. Si, Innovation, 2025, 7, 101048.
26. T. Xu, H. S. Du, H. Y. Liu, W. Liu, X. Y. Zhang, C. L. Si, P. W. Liu and K. Zhang, Adv. Mater., 2021, 33, 2101368.
27. W. S. Chen, H. P. Yu, S. Y. Lee, T. Wei, J. Li and Z. J. Fan, Chem. Soc. Rev., 2018, 47, 2837–2872.
28. W. Liu, K. Liu, H. S. Du, T. Zheng, N. Zhang, T. Xu, B. Pang, X. Y. Zhang, C. L. Si and K. Zhang, Nano-Micro Lett., 2022, 14, 104.
29. W. Li, Y. Xu, K. Liu, L. Y. Zhu, T. Xu, G. H. Wang and C. L. Si, Adv. Mater., 2025, 2503658,  DOI:10.1002/adma.202503658.
30. W. Li, L. Y. Zhu, Y. Xu, G. H. Wang, T. Xu and C. L. Si, Adv. Mater., 2025, 37, 2415761.
31. M. Zhang, T. Xu, K. Liu, L. Y. Zhu, C. Y. Miao, T. Chen, M. G. Gao, J. F. Wang and C. L. Si, Susmat, 2025, 5, e255.
32. Y. F. Yang, L. Yan, Y. Zheng, L. Dai and C. L. Si, Adv. Funct. Mater., 2025, 35, e70571.
33. T. R. Pang, Z. L. Xue, G. H. Wang, J. K. Li, W. J. Sui and C. L. Si, Angew. Chem., Int. Ed., 2025, 64, e202503195.
34. L. Chen, L. Yu, L. H. Qi, S. J. Eichhorn, A. Isogai, E. Lizundia, J. Y. Zhu and C. J. Chen, Nat. Rev. Mater., 2025, 10, 728–779.
35. Q. Dong, X. Zhang, J. Qian, S. M. He, Y. M. Mao, A. H. Brozena, Y. Zhang, T. P. Pollard, O. A. Borodin, Y. B. Wang, B. S. Chava, S. Das, P. Zavalij, C. U. Segre, D. Y. Zhu, L. Xu, Y. L. Liang, Y. Yao, R. M. Briber, T. Li and L. B. Hu, Sci. Adv., 2022, 8, eadd2031.
36. K. Daems, P. Yadav, K. B. Dermenci, J. Van Mierlo and M. Berecibar, Renew. Sustainable Energy Rev., 2024, 191, 114136.
37. Y. T. Zhai, Y. M. Zhang, N. Wang, M. H. Li, M. J. Shen, P. B. Xu, M. Y. Sun, F. L. Dong, T. Z. Ma, J. Ming, L. N. Cong, H. M. Xie and Y. L. Liu, Energy Environ. Sci., 2025, 18, 8352–8365.
38. B. Li, C. H. Wang, R. Z. Yu, J. Q. Han, S. H. Jiang, C. M. Zhang and S. J. He, Energy Environ. Sci., 2024, 17, 1854–1884.
39. W. Liu, S. W. Lee, D. C. Lin, F. F. Shi, S. Wang, A. D. Sendek and Y. Cui, Nat. Energy, 2017, 2, 17035.
40. H. Yang and N. Q. Wu, Energy Sci. Eng., 2022, 10, 1643–1671.
41. J. X. Li, H. Jia, S. S. Ma, L. J. Xie, X. X. Wei, L. Q. Dai, H. L. Wang, F. Y. Su and C. M. Chen, ACS Energy Lett., 2023, 8, 56–78.
42. D. Zhou, D. Shanmukaraj, A. Tkacheva, M. Armand and G. X. Wang, Chem, 2019, 5, 2326–2352.
43. M. J. Li, R. P. Hicks, Z. F. Chen, C. Luo, J. C. Guo, C. S. Wang and Y. H. Xu, Chem. Rev., 2023, 123, 1712–1773.
44. L. Z. Fan, H. C. He and C. W. Nan, Nat. Rev. Mater., 2021, 6, 1003–1019.
45. M. Ahuis, S. Doose, D. Vogt, P. Michalowski, S. Zellmer and A. Kwade, Nat. Energy, 2024, 9, 373–385.
46. J. Lopez, D. G. Mackanic, Y. Cui and Z. N. Bao, Nat. Rev. Mater., 2019, 4, 312–330.
47. L. Y. Zhu, H. B. Yang, T. Xu, F. Shen and C. L. Si, Nano-Micro Lett., 2025, 17, 87.
48. H. Peng, T. R. Long, J. Peng, H. Chen, L. F. Ji, H. Sun, L. Huang and S. G. Sun, Adv. Energy Mater., 2024, 14, 2400428.
49. X. Y. Zhou, Y. F. Zhou, L. Yu, L. H. Qi, K. S. Oh, P. Hu, S. Y. Lee and C. J. Chen, Chem. Soc. Rev., 2024, 53, 5291–5337.
50. T. Janoschka, N. Martin, U. Martin, C. Friebe, S. Morgenstern, H. Hiller, M. D. Hager and U. S. Schubert, Nature, 2015, 527, 78–81.
51. S. H. Zou, Y. Yang, J. R. Wang, X. Y. Zhou, X. H. Wan, M. Zhu and J. Liu, Energy Environ. Sci., 2024, 17, 4426–4460.
52. G. B. Pedersen, L. Blaschek, K. E. H. Frandsen, L. C. Noack and S. Persson, Mol. Plant, 2023, 16, 206–231.
53. Z. M. Wang, S. H. Wang, L. Zhang, H. Liu and X. Xu, Research, 2024, 7, 0298.
54. S. M. Kang, Y. Zhao, J. L. Song and H. N. Xiao, J. Bioresour. Bioprod., 2024, 9, 1–2.
55. C. J. Chen, Y. D. Kuang, S. Z. Zhu, I. Burgert, T. Keplinger, A. Gong, T. Li, L. Berglund, S. J. Eichhorn and L. B. Hu, Nat. Rev. Mater., 2020, 5, 642–666.
56. A. Isogai, Adv. Mater., 2021, 33, 2000630.
57. A. K. Rana, E. Frollini and V. K. Thakur, Int. J. Biol. Macromol., 2021, 182, 1554–1581.
58. Z. Y. Wu, H. W. Liang, L. F. Chen, B. C. Hu and S. H. Yu, Acc. Chem. Res., 2016, 49, 96–105.
59. C. J. Chen, P. F. Xu and X. H. Wang, J. Bioresour. Bioprod., 2024, 9, 102–112.
60. M. Zhang, T. Chen, T. Xu, H. Zhang, X. Wang, J. J. Qi, Q. Dong, L. Y. Zhu, Z. H. Yuan and C. L. Si, Chem. Eng. J., 2025, 506, 159872.
61. D. Trache, M. H. Hussin, M. K. M. Haafiz and V. K. Thakur, Nanoscale, 2017, 9, 1763–1786.
62. J. L. Liu, A. Malvandi and H. Feng, J. Bioresour. Bioprod., 2024, 9, 465–485.
63. F. Rol, M. N. Belgacem, A. Gandini and J. Bras, Prog. Polym. Sci., 2019, 88, 241–264.
64. N. B. Erdal and M. Hakkarainen, Biomacromolecules, 2022, 23, 2713–2729.
65. S. Ummartyotin and H. Manuspiya, Renewable Sustainable Energy Rev., 2015, 41, 402–412.
66. M. Zhou, D. Z. Chen, Q. Q. Chen, P. Chen, G. J. Song and C. Y. Chang, Adv. Mater., 2024, 36, 2312220.
67. Y. J. Deng, T. X. Zhu, Y. Cheng, K. Zhao, Z. Y. Meng, J. Y. Huang, W. L. Cai and Y. K. Lai, Adv. Fiber Mater., 2024, 6, 1343–1368.
68. Q. Qin, S. Y. Zeng, G. G. Duan, Y. B. Liu, X. S. Han, R. Z. Yu, Y. Huang, C. M. Zhang, J. Q. Han and S. H. Jiang, Chem. Soc. Rev., 2024, 53, 9306–9343.
69. J. C. Zhai, W. Zhao, L. Wang, J. B. Shuai, R. W. Chen, W. J. Ge, Y. Zong, G. J. He and X. H. Wang, Energ Environ. Sci., 2025, 18, 4241–4250.
70. X. Yang, M. H. Xie, Z. J. Yan, H. Ruan, C. P. Yang, Z. P. Guo and Z. J. Zheng, Energy Environ. Sci., 2025, 18, 4730–4739.
71. E. J. Foster, R. J. Moon, U. P. Agarwal, M. J. Bortner, J. Bras, S. Camarero-Espinosa, K. J. Chan, M. J. D. Clift, E. D. Cranston, S. J. Eichhorn, D. M. Fox, W. Y. Hamad, L. Heux, B. Jean, M. Korey, W. Nieh, K. J. Ong, M. S. Reid, S. Renneckar, R. Roberts, J. A. Shatkin, J. Simonsen, K. Stinson-Bagby, N. Wanasekara and J. Youngblood, Chem. Soc. Rev., 2018, 47, 2609–2679.
72. Y. H. Ye, L. Yu, E. Lizundia, Y. L. Zhu, C. J. Chen and F. Jiang, Chem. Rev., 2023, 123, 9204–9264.
73. P. Chen, X. H. Lin, B. Yang, Y. Gao, Y. Xiao, L. Li, H. Zhang, L. Li, Z. Zheng, J. Z. Wang and S. L. Chou, Adv. Funct. Mater., 2024, 34, 2409368.
74. H. Yang, L. Zhu, W. Li, Y. Tang, X. Li, T. Xu, K. Liu and C. Si, Nano-Micro Lett., 2026, 18, 84.
75. H. Z. Chen, Z. C. Wang, Y. T. Feng, S. Y. Cai, H. P. Gao, Z. Z. Wei and Y. Zhao, Chem. Eng. J., 2023, 471, 144593.
76. Y. M. Chen, L. Zhang, Y. Yang, B. Pang, W. H. Xu, G. G. Duan, S. H. Jiang and K. Zhang, Adv. Mater., 2021, 33, 2005569.
77. X. Y. Dong, X. Guo, Q. Y. Liu, Y. J. Zhao, H. B. Qi and W. Zhai, Adv. Funct. Mater., 2022, 32, 2203610.
78. W. B. Deng, Y. Liu, C. He, X. Z. Xiong, R. Zhang, T. F. Yan, S. C. Shi, D. G. Yu and H. S. Yang, Chem. Eng. J., 2025, 506, 160117.
79. P. Grzybek, G. Dudek and B. van der Bruggen, Chem. Eng. J., 2024, 495, 153500.
80. J. Hwang, M. H. Myung, J. H. Ha, S. Choi, S. J. Jung, S. Lee, J. Park, Y. R. Kim, H. Y. Jin, N. J. Kong, Y. Kim, H. W. Lee, H. Ko, T. J. Shin, S. J. Kang, M. J. Baek and D. W. Lee, Energy Environ. Sci., 2025, 18, 4447–4459.
81. H. Zhang, J. R. Liang, Z. R. You, Y. Gao, Y. C. Song, D. Lei, C. L. Si and Z. Z. Li, Chem. Eng. J., 2025, 517, 164231.
82. H. Zhang, X. Gan, Z. Wang, M. Tao and J. Zhou, ACS Nano, 2025, 19, 15565–15577.
83. X. Tang, M. Zuo, Z. Li, H. Liu, C. X. Xiong, X. H. Zeng, Y. Sun, L. Hu, S. J. Liu, T. Z. Lei and L. Lin, ChemSusChem, 2017, 10, 2696–2706.
84. W. B. Tu, S. Liang, L. N. Song, X. X. Wang, G. J. Ji and J. J. Xu, Adv. Funct. Mater., 2024, 34, 2316137.
85. L. Shi, G. Wang, J. Li, M. F. Wu and Z. Y. Wen, ACS Sustainable Chem. Eng., 2021, 9, 13883–13892.
86. T. Zhang, X. Q. Zhu, J. Y. Xiong, Z. X. Xue, Y. T. Cao, K. C. Gordon, G. Y. Xu and M. F. Zhu, Nat. Commun., 2025, 16, 4865.
87. W. J. Zhou, M. Yang, M. F. Chen, G. F. Zhang, X. Han, J. Z. Chen, D. T. Ma and P. X. Zhang, Adv. Funct. Mater., 2024, 34, 2315444.
88. J. Li, Z. Hu, S. Zhang, H. Zhang, S. Guo, G. Zhong, Y. Qiao, Z. Peng, Y. Li, S. Chen, G. Chen and A.-M. Cao, Nat. Sustainable, 2024, 7, 1481–1491.
89. R. Wang, W. Dong, Z. Song, J. Tan, Q. Liu, K. Mu, W. Xu, H. Huang, Z. Zhang, G. Yin, C. Zhu, J. Xu and L. Tian, Adv. Funct. Mater., 2024, 34, 2402461.
90. C. A. W. Lu, X. Y. Wang, Y. Shen, C. P. Wang, J. F. Wang, Q. Yong and F. X. Chu, Adv. Funct. Mater., 2022, 32, 2207714.
91. D. Lv, J. C. Chai, P. Wang, L. Y. Zhu, C. Liu, S. X. Nie, B. Li and G. L. Cui, Carbohydr. Polym., 2021, 251, 116975.
92. D. W. Zhao, Y. Zhu, W. K. Cheng, W. S. Chen, Y. Q. Wu and H. P. Yu, Adv. Mater., 2021, 33, 2000619.
93. C. Dahlström, V. L. Durán, S. T. Keene, A. Salleo, M. Norgren and L. Wågberg, Carbohydr. Polym., 2020, 233, 115829.
94. C. P. Yang, Q. S. Wu, W. Q. Xie, X. Zhang, A. Brozena, J. Zheng, M. N. Garaga, B. H. Ko, Y. M. Mao, S. M. He, Y. Gao, P. B. Wang, M. Tyagi, F. Jiao, R. Briber, P. Albertus, C. S. Wang, S. Greenbaum, Y. Y. Hu, A. Isogai, M. Winter, K. Xu, Y. Qi and L. B. Hu, Nature, 2021, 598, 590–596.
95. Y. He, K. Y. Dong, Y. D. Guo, G. Zhao, S. S. Tang, X. L. Wang and B. M. Ma, Polymer, 2023, 287, 126441.
96. C. Yang, P. Woottapanit, S. Geng, R. Chanajaree, Y. Shen, K. Lolupiman, W. Limphirat, T. Pakornchote, T. Bovornratanaraks, X. Zhang, J. Qin and Y. Huang, Nat. Commun., 2025, 16, 183.
97. J. L. Yang, X. X. Zhao, W. Zhang, K. Ren, X. X. Luo, J. M. Cao, S. H. Zheng, W. L. Li and X. L. Wu, Angew. Chem., Int. Ed., 2023, 62, e202300258.
98. L. Xu, T. T. Meng, X. Y. Zheng, T. Y. Li, A. H. Brozena, Y. M. Mao, Q. Zhang, B. C. Clifford, J. C. Rao and L. B. Hu, Adv. Funct. Mater., 2023, 33, 2302098.
99. B. Medronho and B. Lindman, Adv. Colloid Interface Sci., 2015, 222, 502–508.
100. H. Tu, M. X. Zhu, B. Duan and L. N. Zhang, Adv. Mater., 2021, 33, 2000682.
101. J. C. Zhai, W. Zhao, L. Wang, J. B. Shuai, R. W. Chen, W. J. Ge, Y. Zong, G. J. He and X. H. Wang, Energy Environ. Sci., 2025, 18, 4241–4250.
102. X. R. Shi, C. H. Dong, L. Ma, M. R. Yang, X. Chen, Y. Zhang, L. Liu and Z. Long, Chem. Eng. J., 2025, 504, 158659.
103. Y. D. Zhang, W. F. Deng, Z. B. Wang, M. Y. Wu, C. Liu, G. Yu, Q. Li, C. L. Xu and B. Li, ACS Nano, 2025, 19, 21243–21259.
104. Y. Xi, L. L. Zhang, Y. H. Tian, J. L. Song, J. X. Ma and Z. G. Wang, Green Chem., 2022, 24, 885–897.
105. X. D. Liu, Y. C. Tian, L. Wang, L. Chen, Z. P. Jin and Q. Zhang, Nano Lett., 2024, 24, 9074–9081.
106. H. Z. Dou, M. Xu, Z. Zhang, D. Luo, A. P. Yu and Z. W. Chen, Adv. Mater., 2024, 36, 2401858.
107. S. Y. Zhou, S. J. Zhong, Y. F. Dong, Z. Z. Liu, L. W. Dong, B. T. Yuan, H. D. Xie, Y. P. Liu, L. Qiao, J. C. Han and W. D. He, Adv. Funct. Mater., 2023, 33, 2214432.
108. T. Li, X. Zhang, S. D. Lacey, R. Y. Mi, X. P. Zhao, F. Jiang, J. W. Song, Z. Q. Liu, G. Chen, J. Q. Dai, Y. G. Yao, S. Das, R. G. Yang, R. M. Briber and L. B. Hu, Nat. Mater., 2019, 18, 608–613.
109. Y. Lin, S. J. Wang, J. Huang, L. Chen, T. T. Bi, L. H. Qi, Z. Y. Cai, X. Y. Zeng, P. Hu, W. M. Chen, L. Yu and C. J. Chen, Adv. Funct. Mater., 2025, 35, 2504726.
110. C. F. Ding, Y. Liu, L. K. Ono, G. Q. Tong, C. Y. Zhang, J. H. Zhang, J. L. Lan, Y. H. Yu, B. B. Chen and Y. B. Qi, Energy Storage Mater., 2022, 50, 417–425.
111. Y. Yang, J. S. Ding, J. Poisson, R. W. Chen, G. Z. Liu, K. Wang, Y. M. Chen and K. Zhang, ACS Nano, 2025, 19, 15963–15973.
112. X. T. Zhang, J. X. Li, K. W. Qi, Y. Q. Yang, D. Y. Liu, T. Q. Wang, S. Q. Liang, B. A. Lu, Y. C. Zhu and J. Zhou, Adv. Mater., 2022, 34, 2205175.
113. H. C. Zhang, X. Y. Li, J. Hou, L. Jiang and H. T. Wang, Chem. Soc. Rev., 2022, 51, 2224–2254.
114. H. C. Zhang, J. Hou, Y. X. Hu, P. Y. Wang, R. W. Ou, L. Jiang, J. Z. Liu, B. D. Freeman, A. J. Hill and H. T. Wang, Sci. Adv., 2018, 4, eaaq0066.
115. Y. J. Jeon, Y. T. Wu, Y. M. Zhang, C. Hwang, H. W. Lee, H. K. Song and N. A. Liu, Electrochim. Acta, 2021, 391, 138877.
116. Z. J. Xie, Z. R. Xiang, X. T. Fu, Z. W. Lin, C. L. Jiao, K. Zheng, M. Yang, X. Z. Qin and D. D. Ye, ACS Energy Lett., 2024, 9, 2092–2100.
117. X. Y. Wang, Z. Y. Wang, Z. W. Xue, Y. Y. Fan, J. Yang, Q. X. Zhang, N. T. Yang, X. X. Meng, Y. Jin and S. M. Liu, Energy Environ. Sci., 2024, 17, 6225.
118. W. W. Xu, X. B. Liao, W. N. Xu, K. N. Zhao, G. Yao and Q. L. Wu, Adv. Energy Mater., 2023, 13, 2300283.
119. X. Lin, S. Tao, J. Mo, X. Wang, Y. Shao, Y. Hu, C. Qiu, K. Shen, C. Dang and H. Qi, Nat. Commun., 2025, 16, 7527.
120. M. H. Abdellah, C. Oviedo and G. Szekely, J. Membr. Sci., 2023, 687, 122040.
121. N. K. Kwon, T. Kim, S. Yun, H. Kim, H. W. Lee, J. Park and S. Y. Kim, Polymer, 2023, 286, 126401.
122. S. C. Yang, Y. X. Zhang, Y. Zhang, J. Deng, N. X. Chen, S. D. Xie, Y. Ma and Z. H. Wang, Adv. Funct. Mater., 2023, 33, 2304280.
123. Y. T. Zhang, X. Y. Lin, Z. M. Wang, L. Zhang, S. H. Wang, Z. Huang, H. Liu and X. Xu, Chem. Eng. J., 2024, 480, 148318.
124. Z. C. Xu, Y. Q. Yang, X. W. Pang, Y. T. Xu, L. X. Gong, L. C. Tang and S. N. Li, J. Mater. Sci. Technol., 2025, 237, 10–19.
125. M. M. Zhu, J. Y. Yu, Z. L. Li and B. Ding, Angew. Chem., Int. Ed., 2022, 61, e202208949.
126. C. H. Li and J. L. Zuo, Adv. Mater., 2020, 32, 1903762.
127. B. R. Li, P. F. Cao, T. Saito and A. P. Sokolov, Chem. Rev., 2023, 123, 701–735.
128. R. Xue, H. Zhao, Z. W. An, W. Wu, Y. Jiang, P. Li, C. X. Huang, D. Shi, R. K. Y. Li, G. H. Hu and S. F. Wang, ACS Nano, 2023, 17, 5653–5662.
129. F. Han, S. M. Chen, F. Wang, M. Liu, J. H. Li, H. Liu, Y. S. Yang, H. Q. Zhang, D. Liu, R. Y. He, W. T. Cao, X. C. Qin and F. Xu, Adv. Sci., 2025, 12, 2412726.
130. Q. J. Fu, S. W. Hao, X. R. Zhang, H. A. Zhao, F. Xu and J. Yang, Energy Environ. Sci., 2023, 16, 1291–1311.
131. H. X. Wan, Y. Chen, Y. Z. Tao, P. Chen, S. Wang, X. Y. Jiang and A. Lu, ACS Nano, 2023, 17, 20699–20710.
132. Y. Qin, J. L. Mo, Y. H. Liu, S. Zhang, J. L. Wang, Q. Fu, S. F. Wang and S. X. Nie, Adv. Funct. Mater., 2022, 32, 2201846.
133. X. Y. Zhang, Q. W. Sun, C. Zhen, Y. H. Niu, Y. P. Han, G. F. Zeng, D. J. Chen, C. Feng, N. Chen, W. Q. Lv and W. D. He, Energy Storage Mater., 2021, 37, 628–647.
134. D. X. Li, Y. Ouyang, Y. B. Xiao, Y. F. Xie, Q. H. Zeng, S. T. Yu, C. Zheng, Q. Zhang and S. M. Huang, Adv. Funct. Mater., 2024, 34, 2314296.
135. J. Y. Ren, J. X. Guo, X. Y. Luo, X. Y. Dai, G. W. Zhang, P. Ping and D. P. Kong, J. Power Sources, 2025, 628, 235862.
136. Z. Wang, P. Heasman, J. Rostami, T. Benselfelt, M. Linares, H. L. Li, A. Iakunkov, F. Sellman, R. Östmans, M. M. Hamedi, I. Zozoulenko and L. Wågberg, Adv. Funct. Mater., 2023, 33, 2212806.
137. Y. H. Liu, Y. Y. Yu, Y. T. Liu, S. Y. Li, C. Z. Liu, W. Hu, Y. F. Lu and G. S. Zhu, Adv. Energy Mater., 2025, 15(33), 2501737.
138. Z. Q. Zhang, G. Li, X. F. Du, L. Huang, G. H. Kang, J. J. Zhang, Z. L. Cui, T. Liu, L. Ni, Y. C. Jin and G. L. Cui, Adv. Sci., 2024, 11, 2409628.
139. X. F. Zi, H. M. Wu, J. L. Song, J. Q. Wu and J. B. Guo, ACS Appl. Mater. Interfaces, 2024, 16, 47416–47428.
140. A. Permatasari, M. G. So, V. Nguyen, T. Yano and G. Inoue, J. Power Sources, 2025, 647, 237278.
141. D. W. Zhao, C. J. Chen, Q. Zhang, W. S. Chen, S. X. Liu, Q. W. Wang, Y. X. Liu, J. Li and H. P. Yu, Adv. Energy Mater., 2017, 7, 1700739.
142. X. Y. Wang, W. Q. Zheng, H. Zhao, J. Y. Li, S. Chen and F. Xu, Nano-Micro Lett., 2025, 17, 153.
143. X. M. Hu, G. Y. Han, Y. R. Deng, Z. Y. Yang, X. X. Wei, H. Y. Xu and Z. J. Zhang, J. Colloid Interface Sci., 2025, 679, 633–642.
144. T. H. Wang, Q. H. Liu, J. Zhou, X. H. Wang and B. A. Lu, Adv. Energy Mater., 2022, 12, 2202357.
145. T. C. Mokhena, E. R. Sadiku, M. J. Mochane, S. S. Ray, M. J. John and A. Mtibe, Carbohydr. Polym., 2021, 273, 118507.
146. C. Cheng, R. D. Yang, Y. Wang, X. Y. Dong, D. N. Fu, J. Sheng and X. H. Guo, Int. J. Biol. Macromol., 2024, 283, 137407.
147. X. Gong, J. Wang, L. Zhong, G. Qi, F. Liu, Y. Pan, F. Yang, X. Wang, J. Li, L. Li, C. Liu and D. Yu, Ind. Chem. Mater., 2025, 3, 31–48.
148. F. Chen, X. Li, S. P. Chen, Y. L. Zhang, H. D. Huang, H. L. Yang, S. Y. Zhou and Z. M. Li, Adv. Sci., 2025, 12, e11759.
149. N. G. Park, Mater. Today, 2015, 18, 65–72.
150. T. Leijtens, K. A. Bush, R. Prasanna and M. D. McGehee, Nat. Energy, 2018, 3, 828–838.
151. H. W. Zhu, S. Teale, M. N. Lintangpradipto, S. Mahesh, B. Chen, M. D. McGehee, E. H. Sargent and O. M. Bakr, Nat. Rev. Mater., 2023, 8, 569–586.
152. Q. X. Fu, X. L. Tang, B. Huang, T. Hu, L. C. Tan, L. Chen and Y. W. Chen, Adv. Sci., 2018, 5, 1700387.
153. F. Bisconti, A. Giuri, L. Dominici, S. Carallo, E. Quadrivi, R. Po, P. Biagini, A. Listorti, C. E. Corcione, S. Colella and A. Rizzo, Nano Energy, 2021, 89, 106406.
154. J. M. Yang, S. B. Xiong, T. Y. Qu, Y. X. Zhang, X. X. He, X. W. Guo, Q. H. Zhao, S. Braun, J. Q. Chen, J. H. Xu, Y. Q. Li, X. J. Liu, C. G. Duan, J. X. Tang, M. Fahlman and Q. Bao, ACS Appl. Mater. Interfaces, 2019, 11, 13491–13498.
155. X. F. Zhu, R. Lin, H. Gu, H. C. Hu, Z. Liu, G. C. Xing, Y. B. Wu and X. H. Ouyang, Energy Environ. Mater., 2023, 6, e12426.
156. Z. L. Zhang, C. Wang, F. Li, L. S. Liang, L. L. Huang, L. H. Chen, Y. H. Ni, P. Gao and H. Wu, Adv. Sci., 2023, 10, 2207202.
157. Z. L. Zhang, C. M. Liu, C. Wang, J. D. Liang, L. S. Liang, X. Z. Huang, D. D. Song, Y. Wang, L. L. Huang, L. H. Chen, A. R. B. Yusoff, P. Gao and H. Wu, Sci. China Chem., 2025, 68(10), 5176–5186.
158. D. Lu, Z. Fang and Z. H. Qiao, Nano Lett., 2025, 25, 2850–2857.
159. Q. W. Lu, Y. B. He, Q. P. Yu, B. H. Li, Y. V. Kaneti, Y. W. Yao, F. Y. Kang and Q. H. Yang, Adv. Mater., 2017, 29, 1604460.
160. Z. Y. Song, F. F. Chen, M. Martinez-Ibañez, W. F. Feng, M. Forsyth, Z. B. Zhou, M. Armand and H. Zhang, Nat. Commun., 2023, 14, 4884.
161. F. Zhao, J. H. Zhang, J. X. Chen, Z. Y. Gu, X. Z. Fan, L. Zhu, H. L. Na, M. X. Dong, C. Guan and L. Kong, Research, 2025, 8, 0802.
162. P. P. Dong, Y. H. Cha, X. H. Zhang, J. Zamora and M. K. Song, ACS Appl. Mater. Interfaces, 2024, 16, 41018–41026.
163. Y. J. Li, T. T. Gao, D. Y. Ni, Y. Zhou, M. Yousaf, Z. Q. Guo, J. H. Zhou, P. Zhou, Q. Wang and S. J. Guo, Adv. Mater., 2022, 34, 2107638.
164. Z. P. Liu, B. Sun, Y. Zhang, Q. X. Zhang and L. S. Fan, Prog. Polym. Sci., 2024, 152, 101817.
165. K. Wu, J. H. Huang, J. Yi, X. Y. Liu, Y. Y. Liu, Y. G. Wang, J. J. Zhang and Y. Y. Xia, Adv. Energy Mater., 2020, 10, 1903977.
166. C. Q. Chao, Y. N. Li and Y. Zhao, Rare Met., 2024, 43, 4807–4843.
167. Q. H. Wang, J. Huang, L. H. Qi, M. Li, S. J. Wang, J. Q. Chen, Z. Y. Sui, T. T. Bi, Q. C. Tang, L. Yu, P. Hu, W. Zhang, C. H. Lu and C. J. Chen, ACS Nano, 2025, 19, 26770–26781.
168. M. Zhang, T. Xu, W. Liu, H. Zhang, J. J. Qi, X. Wang, Y. X. Wang, L. Y. Zhu, K. Liu, J. F. Wang and C. L. Si, Carbon Energy, 2025, e70097,  DOI:10.1002/cey2.70097.
169. Q. S. Zhao, T. Xu, K. Liu, H. S. Du, M. Zhang, Y. X. Wang, L. X. Yang, H. Zhang, X. Wang and C. L. Si, Energy Storage Mater., 2024, 71, 103605.
170. M. Y. Liu, L. M. Zhang, J. Rostami, T. Zhang, K. Matthews, S. Chen, W. J. Fan, Y. Zhu, J. W. Chen, M. H. Huang, J. Y. Wu, H. L. Wang, M. M. Hamedi, F. Xu, W. Q. Tian, L. Wågberg and Y. Gogotsi, ACS Nano, 2025, 19, 13399–13413.
171. Z. C. Bai, G. L. Wang, H. M. Liu, Y. T. Lou, N. A. Wang, H. K. Liu and S. X. Dou, Chem. Sci., 2024, 15, 3071–3092.
172. J. L. Yang, T. Xiao, T. Xiao, J. Li, Z. H. Yu, K. Liu, P. H. Yang and H. J. Fan, Adv. Mater., 2024, 36, 2313610.
173. J. Huang, S. J. Wang, L. Yu, E. Lizundia, X. Y. Zeng, L. Chen, Z. Y. Lu, L. H. Qi, Y. Lin, H. B. Deng and C. J. Chen, Natl. Sci. Rev., 2025, 12, nwaf308.
174. J. Yang, G. Z. Liu, M. Avdeev, H. L. Wan, F. D. Han, L. Shen, Z. Y. Zou, S. Q. Shi, Y. S. Hu, C. S. Wang and X. Y. Yao, ACS Energy Lett., 2020, 5, 2835–2841.
175. Y. Q. Li, L. Y. Zhang, X. L. Wang, X. H. Xia, D. Xie, C. D. Gu and J. P. Tu, Research, 2019, 2019, 6930294.
176. X. S. Li, J. Liang, X. Cao, S. Y. Zhu, Y. F. Bai, J. W. Sun, H. B. Luo and J. Kong, Rare Met., 2025, 44, 2871–2899.
177. H. C. Gao, S. Xin, L. G. Xue and J. B. Goodenough, Chem, 2018, 4, 833–844.
178. D. Zhou, X. Tang, X. Guo, P. Li, D. Shanmukaraj, H. Liu, X. C. Gao, Y. Z. Wang, T. Rojo, M. Armand and G. X. Wang, Angew. Chem., Int. Ed., 2020, 59, 16725–16734.
179. Z. X. Xue, T. Zhang, X. Li, F. Wang, G. Y. Xu and M. F. Zhu, Angew. Chem., Int. Ed., 2025, 64, e202415283.
180. X. F. Xu, D. Zhou, X. Y. Qin, K. Lin, F. Y. Kang, B. H. Li, D. Shanmukaraj, T. Rojo, M. Armand and G. X. Wang, Nat. Commun., 2018, 9, 3870.
181. Y. J. Lei, X. X. Lu, H. Yoshikawa, D. Matsumura, Y. M. Fan, L. F. Zhao, J. Li, S. Wang, Q. F. Gu, H. K. Liu, S. X. Dou, S. Devaraj, T. Rojo, W. H. Lai, M. Armand, Y. X. Wang and G. X. Wang, Nat. Commun., 2024, 15, 3325.
182. S. M. Luo, J. Shang, Y. N. Xu, H. Cheng, L. J. Zhang and Y. B. Tang, Adv. Funct. Mater., 2024, 34, 2403166.
183. W. Yang, W. Yang, R. Zou, Y. F. Huang, H. H. Lai, Z. H. Chen and X. W. Peng, Carbon Energy, 2023, 5, e203.
184. H. C. Yang, H. C. Li, J. Li, Z. H. Sun, K. He, H. M. Cheng and F. Li, Angew. Chem., Int. Ed., 2019, 58, 11978–11996.
185. G. A. Elia, K. Marquardt, K. Hoeppner, S. Fantini, R. Y. Lin, E. Knipping, W. Peters, J. F. Drillet, S. Passerini and R. Hahn, Adv. Mater., 2016, 28, 7564–7579.
186. Y. Zhang, S. Q. Liu, Y. J. Ji, J. M. Ma and H. J. Yu, Adv. Mater., 2018, 30, 1706310.
187. Y. F. Liu, Y. H. Xie, X. Wang, L. L. Xiong, D. Li, S. X. Wang, X. Li, W. Z. You, L. H. Cheng, L. Li and A. Q. Zheng, Nano Energy, 2024, 130, 110140.
188. Y. S. Liu, Q. Sun, W. Z. Li, K. R. Adair, J. Li and X. L. Sun, Green Energy Environ., 2017, 2, 246–277.
189. R. Mori, Electrochem. Energy Rev., 2020, 3, 344–369.
190. Y. F. Wang, H. Y. H. Kwok, W. D. Pan, Y. G. Zhang, H. M. Zhang, X. Lu and D. Y. C. Leung, Electrochim. Acta, 2019, 319, 947–957.
191. F. X. Wang, X. W. Wu, X. H. Yuan, Z. C. Liu, Y. Zhang, L. J. Fu, Y. S. Zhu, Q. M. Zhou, Y. P. Wu and W. Huang, Chem. Soc. Rev., 2017, 46, 6816–6854.
192. K. Keum, J. W. Kim, S. Y. Hong, J. G. Son, S. S. Lee and J. S. Ha, Adv. Mater., 2020, 32, 2002180.
193. H. Y. Liu, T. Xu, Q. D. Liang, Q. S. Zhao, D. W. Zhao and C. L. Si, Adv. Compos. Hybrid Mater., 2022, 5, 1168–1179.
194. Y. Y. Wang, S. R. Sun, X. L. Wu, H. F. Liang and W. L. Zhang, Nano-Micro Lett., 2023, 15, 78.
195. Q. H. Wu, T. Q. He, Y. K. Zhang, J. L. Zhang, Z. J. Wang, Y. Liu, L. Zhao, Y. Z. Wu and F. Ran, J. Mater. Chem. A, 2021, 9, 24094–24147.
196. Q. D. Liang, K. Liu, T. Xu, Y. X. Wang, M. Zhang, Q. S. Zhao, W. R. Zhong, X. M. Cai, Z. J. Zhao and C. L. Si, Small, 2024, 20, 2470158.
197. Z. C. Ling, Q. He, H. B. Yang, Z. Zhou, Z. M. Han, X. H. Luo, K. P. Yang, Q. F. Guan and S. H. Yu, Adv. Mater., 2024, 36, 2402695.
198. C. J. Chen, Y. Zhang, Y. J. Li, J. Q. Dai, J. W. Song, Y. G. Yao, Y. H. Gong, I. Kierzewski, J. Xie and L. B. Hu, Energy Environ. Sci., 2017, 10, 538–545.
199. N. Raza, K. Aziz, M. S. Javed, R. Tariq, N. Kanwal, W. Raza, M. Sarfraz, S. Khan, M. A. Ismail, B. Akkinepally and X. Wang, J. Energy Storage, 2025, 111, 115328.
200. X. L. Li, L. B. Yuan, R. Liu, H. N. He, J. N. Hao, Y. Lu, Y. M. Wang, G. M. Liang, G. H. Yuan and Z. P. Guo, Adv. Energy Mater., 2021, 11, 2003010.
201. L. Y. Zhu, H. B. Yang, T. Xu, L. Y. Wang, J. D. Lei and C. L. Si, Adv. Funct. Mater., 2025, 35, 2419334.
202. F. Yang, G. Xu, Y. B. Dou, B. Wang, H. Zhang, H. Wu, W. Zhou, J. R. Li and B. L. Chen, Nat. Energy, 2017, 2, 877–883.
203. L. Y. Zhu, L. M. Zhang, Y. T. Ren, J. D. Lei, L. Y. Wang and J. Liu, Adv. Funct. Mater., 2024, 34, 2313844.
204. L. Y. Zhu, P. Ye, L. M. Zhang, Y. T. Ren, J. Liu, J. D. Lei and L. Y. Wang, Small, 2024, 20, 2304575.
205. L. Y. Zhu, Y. C. Li, P. Ye, J. Y. Zhao, J. Liu, J. D. Lei, L. Y. Wang and R. S. Xue, Adv. Funct. Mater., 2023, 33, 2210453.
206. C. Huang, G. L. Zhao, Y. H. Song, C. Xie, S. H. Zhang and X. F. Li, ACS Sustainable Chem. Eng., 2022, 10, 5559–5568.
207. X. Cheng, J. C. Wang, Y. C. Liao, C. P. Li and Z. D. Wei, ACS Appl. Mater. Interfaces, 2018, 10, 23774–23782.
208. D. Yang, R. X. Yang, S. Priya and S. Z. Liu, Angew. Chem., Int. Ed., 2019, 58, 4466–4483.
209. Y. N. Jiao, Y. Wang, Z. X. Shang, Y. C. Liang, K. Y. Sun, W. W. Wang, S. H. Yi, Z. L. Wang, J. X. Guo, M. G. Ma, D. J. Dong, M. X. Wu and J. J. Zhao, Rare Met., 2025, 44, 1717–1729.
210. Q. Jiang, Z. N. Chu, P. Y. Wang, X. L. Yang, H. Liu, Y. Wang, Z. G. Yin, J. L. Wu, X. W. Zhang and J. B. You, Adv. Mater., 2017, 29, 1703852.
211. Y. R. Wang, Y. H. Cheng, C. C. Yin, J. M. Zhang, J. X. You, J. Z. Wang, J. F. Wang and J. Zhang, Nano-Micro Lett., 2024, 16, 183.
212. W. W. Xin, Z. Zhang, X. D. Huang, Y. H. Hu, T. Zhou, C. C. Zhu, X. Y. Kong, L. Jiang and L. P. Wen, Nat. Commun., 2019, 10, 3876.
213. R. A. Tufa, S. Pawlowski, J. Veerman, K. Bouzek, E. Fontananova, G. di Profio, S. Velizarov, J. G. Crespo, K. Nijmeijer and E. Curcio, Appl. Energy, 2018, 225, 290–331.
214. C. C. Zhu, P. Liu, B. Niu, Y. N. Liu, W. W. Xin, W. P. Chen, X. Y. Kong, Z. Zhang, L. Jiang and L. P. Wen, J. Am. Chem. Soc., 2021, 143, 1932–1940.
215. N. Sheng, M. H. Zhang, Q. Song, H. Zhang, S. Y. Chen, H. P. Wang and K. Zhang, Nano Energy, 2022, 101, 107548.
216. J. Huang, S. J. Wang, J. Q. Chen, C. J. Chen and E. Lizundia, Adv. Mater., 2025, 37, 2416733.
217. Y. E. Kim, S. J. Yeom, J. E. Lee, S. J. Kang, H. S. Kang, G. H. Lee, M. J. Kim, S. G. Lee, H. W. Lee and H. G. Chae, J. Power Sources, 2020, 468, 228371.
218. Y. Cho, P. T. T. Ninh, S. Hwang, S. Choe and J. Myung, ACS Mater. Lett., 2025, 7, 1563–1592.

---