Phosphorylation-assisted cell wall engineering enables ultra-strong, highly ion-conductive bio-membranes for high-power salinity gradient energy harvesting†

Kaihuang Chen‡ ^ab, Jie Zhou‡^c, Chunbao Charles Xu^b, Zhiqiang Fang*^a, Le Yu^c, Chaoji Chen*^c and Xueqing Qiu*^d
aState Key Laboratory of Advanced Papermaking and Paper-based Materials, School of Light Industry and Engineering, South China University of Technology, Guangzhou 510640, Guangdong, P. R. China. E-mail: mszhqfang@scut.edu.cn
^bSchool of Energy and Environment, City University of Hong Kong, Hong Kong SAR, China
^cHubei Biomass-Resource Chemistry and Environmental Biotechnology Key Laboratory, Hubei Provincial Engineering Research Center of Emerging Functional Coating Materials, School of Resource and Environmental Sciences, Wuhan University, Wuhan, 430079, China. E-mail: chenchaojili@whu.edu.cn
^dSchool of Chemical Engineering and Light Industry, Guangdong University of Technology, Waihuan Xi Road 100, Guangzhou 510006, China. E-mail: cexqqiu@scut.edu.cn

First published on 17th July 2025

---

---

New concepts Nanofluidic membranes derived from natural materials hold great promise for ion transport and regulation but struggle with the trade-off between mechanical strength and ionic conductivity. Here, we introduce phosphorylation-assisted cell wall engineering, achieving a balance between multiscale structural preservation and high charge density. The resulting phosphorylated bio-membrane exhibits exceptional tensile strength (819.8 MPa) and record-high ionic conductivity (21.01 mS cm−1). Furthermore, cation selectivity enhanced by high charge density facilitates efficient salinity-gradient energy harvesting, outperforming reported cellulose-based nanofluidic systems. This work provides a novel and efficient method to leverage renewable biomaterials for developing high-performance nanofluidic systems.

Introduction

Nanofluidic membranes have aroused great research interest due to their selective ion transportation and ultra-high conductivity at low ion concentrations,¹ which make them promising for applications in nanofluidic energy capture,² sensors,^3,4 ultrafiltration,⁵ biochemistry,^6,7 seawater desalination,⁸ and energy storage.^9,10 The key factors influencing ion transport within these membranes are the charge density and the structural configuration of the ion channels, both of which critically determine ion mobility and selectivity.^11,12 Unlike in bulk systems, the ion conductivity in nanofluidic channels is primarily governed by the concentration of mobile counterions. This is heavily influenced by the effective charge density on the channel surface and the bulk salt concentration dictated by the electrolyte concentrations within the channel.^13,14 To achieve ultra-high ionic conductivity and selectivity at low salt concentrations, it is essential to minimize the contribution from the bulk portion-typically by decreasing thickness to less than 100 nm-and increase the effective charge density.^15,16

Over the years, various materials and fabrication methods have been explored to create high-performance nanofluidic devices with well-aligned structure and improved surface charge density.¹⁷ Among these, the bottom-up approach has gained popularity, particularly using nanosheets (e.g., GO,^18,19 C₃N₄,²⁰ BN,²¹ MMT,²² VMT,²³ MoS₂,^24,25 V₂O₅,²⁶ Ti₃C₂,²⁷ CONs²⁸) and/or polymers (e.g., Kevlar,²⁹ PVDF,³⁰ cellulose,^31,32 and ionomer resin³³) as building blocks. However, this approach faces several challenges. While nanosheets contribute to the formation of internal laminar structures, their limited surface functionalization hampers chemical modification to enhance surface charge density.^34–36 Additionally, the inferior mechanical strength and flexibility of nanosheet-based nanofluidic membranes restrict their practical applications.³⁷ Synthetic polymers offer advantages in addressing these issues, yet they often require expensive processes such as like wet-stretching,³⁸ wet-extrusion³⁹ and framework combinations⁴⁰ to obtain aligned ion channel structures. Even when combining nanosheets with polymers, issues such as low charge density, insufficient mechanical strength, and excessive swelling continue to pose significant hurdles.^41–43 Therefore, there remains an urgent need to develop nanofluidic membranes with both high charge density and well-ordered structures through efficient and straightforward fabrication methods.

While researchers strive to explore methods for constructing ordered structures to regulate ion transport, hierarchical aligned structures that facilitate ion translocation have evolved naturally in trees over hundreds of millions of years.⁴⁴ As depicted in Fig. 1(A), trees effectively utilize transpiration to transport ions from their roots to leaves, spanning distances of over one hundred meters. This process relies on highly aligned cellulose-based structures that not only enable the efficient movement of water and nutrients but also provide the necessary mechanical support to the tree, allowing it to thrive under natural environmental conditions. These unique properties of cellulose make it an ideal material for constructing nanofluidic membranes, offering both structural integrity and ionic conductivity.^45,46 Researchers have employed various methods to modify wood for nanofluidic applications, including both physical methods (e.g., densification,⁴⁷ inner coating⁴⁸ and lumen filling⁴⁹) and chemical approaches (e.g., delignification,⁵⁰ functional group grafting⁵¹ and hydrogel infiltrating⁵²). However, these conventional modification strategies often lead to compromises, such as low charge density and significant disruption of the aligned cellulose structures.^53,54 These issues result in suboptimal ion-conducting performance and weakened mechanical properties, which limit the potential of these bio-based nanofluidic membranes.

In our design, we address these challenges by a phosphorylation-assisted cell wall engineering strategy. By carefully controlling the introduction of phosphate groups (–PO₄²⁻), a perfect balance between high charge density and multiscale structural alignment preservation is achieved (Fig. 1(A)). At the molecular scale, phosphorylation significantly enhances the charge density by introducing anionic phosphate groups onto cellulose chains. Concurrently, at larger scales, this process preserves the native fibril alignment and macro-structural integrity of cellulose, ensuring a hierarchical organization from molecular to macroscopic levels. Following a mechanical press at 10 MPa and 90 °C, the resulting PhosWood-40 membranes exhibit a record-high ion conductivity of 21.01 mS cm⁻¹ in 1.0 × 10⁻⁵ mol L⁻¹ KCl aqueous solution, an ionic selectivity of 0.95, and a tensile strength of up to 241 MPa under dry conditions, as shown in Fig. 1(B). Additionally, the hosWood-40 membranes demonstrate outstanding salinity-gradient power conversion performance of 6.4 W m⁻², which significantly outperforms other ionic nanofluidic membranes (Fig. 1(C)). Unlike previous studies that focused on either charge density or structural alignment, our work achieves a perfect balance between high charge density and multiscale structural alignment, resulting in unprecedented ion conductivity and mechanical strength. These findings offer a promising avenue for the application of wood nanofluidic membranes in advanced ion-conducting devices and provide valuable guidance for the continued development of bio-based nanofluidic membranes.

Results and discussion

Chemical characterization of phosphorylated bio-membranes

A phosphorylation-assisted cell wall engineering strategy was adopted to prepare ultra-strong and highly ion-conductive bio-membranes. Phosphorylation was initially employed to chemically modify the cell walls of delignified pinewood by introducing negative charge groups onto the cellulose molecular chain, owing to its high modification rate and efficiency.⁵⁵ Fig. 2(A) shows a schematic illustration of this phosphorylation process. Cellulose was phosphorylated using diammonium hydrogen phosphate dissolved in molten urea as a solvent at 170 °C for different durations. During this process, the hydroxyl groups of cellulose chemically react with phosphate salts, forming a high density of negatively charged phosphate groups (–PO₄²⁻).⁵⁶ Fourier transform infrared (FTIR) spectroscopy, nuclear magnetic resonance (NMR), and X-ray photoelectron spectroscopy (XPS) were employed to characterize the chemical structure of phosphorylated cellulose in bio-membranes at different phosphorylation durations. As shown in Fig. 2(B), peaks in the range of 3200–3800 cm⁻¹, corresponding to the –OH stretching at the C₂, C₃ and C₆ positions of native cellulose, undergo a blueshift and exhibit increased intensity with prolonged phosphorylation, suggesting the introduction of new hydrogen bonds. While peaks around 2800 cm⁻¹ represent –CH₂– stretching.⁵⁷ Peaks at 800–1000 cm⁻¹ and 1300 cm⁻¹, attributed to P–O– and PO stretching vibrations in the phosphate groups, respectively, appear and intensify progressively, confirming the successful grafting of phosphate groups onto cellulose chains.

As shown in Fig. 2(C), the ¹³C nuclear magnetic resonance (NMR) spectra indicates that the molecular structure of the cellulose remains largely intact after phosphorylation. The characteristic peaks of cellulose, including those corresponding to C2, C3, C5 carbons in the 69–79 ppm region, and C1 carbons in the 101–108 ppm region, are preserved. Compared to natural wood, PhosWood-0 (unmodified bio-membrane) exhibits typical cellulose peaks in the 69–79 ppm region for C2, C3, and C5 carbons, in the 82–90 ppm region for C4, in the 60–67 ppm region for C6 and in the 101–108 ppm region for C1. For PhosWood-40, subtle shifts in these regions are observed, yet the data confirm the preservation of the bulk chemical structure following phosphorylation, despite potential changes in the crystalline structure of cellulose.⁵⁸ Moreover, the total peak area attributable to the two C4 carbon domains in each sample is almost identical to the total peak area associated with the C1 carbon domain, indicating no regional selectivity. This observation suggests that phosphate grafting is uniformly distributed along the cellulose fibers, leading to no specific changes in the ¹³C chemical shifts. X-ray photoelectron spectroscopy (XPS) results complement the FTIR findings, providing further confirmation of successful phosphorylation. In the full-scan XPS spectra (Fig. S2, ESI†), the PhosWood-0 exhibits only C1s and O1s signals, whereas phosphorylated membranes show additional N1s, P2s, and P2p signals at binding energies of 401, 191, and 134 eV, respectively. High-resolution XPS spectra of O1s and P2p (Fig. 2(D) and (E)) reveal increased signal intensity with extended phosphorylation duration, indicating higher phosphate group incorporation.⁵⁹ These observations confirm the successful introduction of phosphate groups into the cellulose, with the degree of phosphate substitution increasing progressively with reaction time (Fig. 2(F)). Phosphate substitution reaches a peak value of 1.411 after 40 minutes of reaction, while further extension to 60 minutes leads to only a marginal increase to 1.461. This suggests that most reactive hydroxyl sites have already been functionalized at 40 minutes, and the phosphorylation reaction approaches saturation thereafter.

Optical microscopy (Fig. S3C–G, ESI†) further reveals significant swelling of cellulose fibers with increasing phosphorylation, indicating the effective opening of ion channels. This structural evolution enhances water absorption and improves ion transport capacity, making the phosphorylated bio-membrane highly suitable for ion-selective applications. Water absorption rises from 7.9% in PhosWood-0 to 23.3% in PhosWood-40 (Fig. 2(H)), but shows negligible further increase in PhosWood-60 (23.2%), supporting the conclusion that hydrophilic functionalization saturates beyond 40 minutes. This enhancement is attributed to disruption of crystalline cellulose regions and fiber swelling, which increase ion channel accessibility and facilitate water penetration. X-Ray diffraction (XRD) and optical microscopy were used to characterize these structural changes (Fig. 2(G) and Fig. S3, ESI†). All membranes exhibit characteristic peaks of type I cellulose crystals at around 23° and 15°. However, crystallinity decreases with increasing phosphorylation. Specially, the crystallinity drops from 76.7% in PhosWood-0 to 67.8% in PhosWood-60, indicating a gradual breakdown of the crystalline region during phosphorylation. These morphological changes, coupled with increased channel accessibility, enable greater water penetration and improved ion transport in the phosphorylated bio-membranes. The morphological changes in cellulose fibers dispersed in water also correlate with the degree of phosphorylation. With higher phosphorylation levels, the cellulose fibers exhibit noticeable swelling (Fig. S3C–G, ESI†), suggesting sufficient opening of the ion channels. These morphological changes, coupled with increased channel accessibility, enable greater water penetration and improved ion transport in the phosphorylated bio-membranes.

To evaluate the structural variations between natural wood and phosphorylated bio-membranes at different hierarchical levels, we investigated the morphology of those membranes at the macroscopic, micro, nano, and molecular scales (Fig. 3). Comparative observations using digital photographs, optical microscopy, and scanning electron microscopy (SEM) images show that all samples maintain highly aligned configurations along the direction of tree growth, consistent with natural wood at the macroscopic, micro, and nano levels (Fig. S4, ESI†), which remained identical to natural wood. SEM images in Fig. 3(D) exhibit the typical top view of natural pinewood, the cross-section contains numerous rectangular holes (cell lumina of ∼20 μm in diameter). This porous structure was eliminated by physical densification treatment in all phosphorylated bio-membranes. The transmittance test further supported these observations. PhosWood-0 shows the highest transmittance, with 72.3% transmittance at 600 nm, and the text behind the membranes remains clearly visible, indicating a dense and defect-free internal structure (Fig. S5 and S6, ESI†). However, as the phosphorylation duration increases, the transmittance of the phosphorylated bio-membranes decreases. Specifically, the transmittance at 600 nm drops from 69.0% for PhosWood-5 to 40.9% for PhosWood-60, which is due to the increase of voids and defective structures within the membranes.

Small-angle X-ray scattering (SAXS) further confirms that both dry and wet phosphorylated bio-membranes (Fig. 3(M)–(Q)) retain molecular-level alignment consistent with the natural wood structure. From SAXS scattering intensity versus azimuthal angle curves (Fig. 3(R)), the degree of orientation (Π) was calculated based on the full width at half-maximum (marked by the dotted line). The orientation degree of all the membranes exceeds 0.5, indicating a high level of orientation parallel to the tree growth direction. Notably, the orientation degree of the PhosWood-40 membrane remains stable in both dry and swollen states (∼0.8), indicating the stability of the ordered structure in water. Although the orientation degree of PhosWood-40 membrane is slightly lower than that of PhosWood-0 membrane due to its internal defects, it remains higher than that of natural wood, which has an orientation degree of 0.69. This suggests that the phosphorylation-assisted cell wall engineering strategy effectively enhances the structural merits of natural wood. To determine the morphological evolution of cellulose nanofibers within phosphorylated bio-membranes affected by phosphorylation modification and water absorption, we investigated the average distance between cellulose nanofibers and the average diameter of cellulose nanofibers. SAXS profiles of various phosphorylated bio-membranes were obtained through the integration of 2D SAXS patterns, and the fitting curves were derived using the WoodSAS model⁶⁰ (Fig. 3(S)). In all dry membranes, the crystalline domains of cellulose nanofiber exhibited a highly oriented microstructure as indicated in Fig. 3(T), the average distance (∼2.9 nm) and diameter (∼2.0 nm) of cellulose nanofibers remain essentially unchanged. During the water absorption process of the membranes, the diameter of the cellulose nanofibers remained almost unchanged, but the average distance between cellulose nanofibers increased from 2.90 nm in dry membranes to 5.83 nm in confined swelling PhosWood-40, and further to 7.56 nm when PhosWood-40 is fully swollen. The highly ordered structure remained intact, while the average spacing of the cellulose nanofibers increased, which may significantly enhance the ionic conductivity of the membranes.

Tensile strength and ion conductivity of phosphorylated bio-membranes

The phosphorylated bio-membranes exhibit both ultra-high ion conductivity and good tensile strength, attributed to the introduction of abundant phosphate groups and the preservation of hierarchical aligned wood structure. Fig. 4(A) and Fig. S8 (ESI†) shows the stress–strain curves of phosphorylated bio-membranes, while their tensile strength and modulus are shown in Fig. 4(B). The PhosWood-0 exhibits the highest strength (819 MPa), elongation (7.1%) and modulus (7.1 GPa). Although both the strength and modulus of the phosphorylated bio-membranes decrease with the increase of phosphorylation degree, the PhosWood-40 membrane still exhibits tensile strength up to 241 MPa, which is twice as strong as natural pinewood. To evaluate the ionic transport properties of the membranes, impedance tests were performed in different concentrations of electrolyte solutions, as shown in Fig. 4(C) (Experimental section). Fig. 4(D) shows typical Nyquist plots of various phosphorylated bio-membranes in a 10⁻⁴ mol L⁻¹ KCl solution. The AC conductivity test results shown in Fig. S9 (ESI†) indicate that the conductivity tested by ESI† is only contributed by ion transport. By measuring the resistance of the nanofluidic devices in different KCl concentrations, the ionic conductivity of the membranes is calculated (Fig. 4(E)). All membranes exhibit typical nanofluidic ionic conductivity behavior. The ion conductivity remains consistent at low KCl concentrations (below 10⁻³ mol L⁻¹), and, as expected, the ion conductivity of phosphorylated bio-membranes is enhanced with the increase of phosphorylation. At KCl concentrations above 10⁻³ mol L⁻¹, the ion conductivity of all membranes increases with the increase of concentration due to the gradual dominance of the salt ions in the bulk solution. The differences in ionic conductivity between membranes can be attributed to the varying ion channel densities. At a KCl concentration of 10⁻⁴ mol L⁻¹, PhosWood-0 exhibited an ion conductivity of 0.21 mS cm⁻¹, while PhosWood-40 showed an ionic conductivity approximately 100 times higher than that of PhosWood-0, reaching 21.01 mS cm⁻¹. This significant enhancement is primarily due to the concentration of mobile counterions, which is governed by the effective surface charge density on the ion channels. As PhosWood-40 membranes contain more charged ions on the surface of the ion channels, they demonstrate superior ionic conductivity at lower KCl concentrations. The negatively charged phosphate groups in water facilitate cation transport, leading to higher ionic conductivity and cation selectivity with increased phosphorylation. This is confirmed by the charge densities of the membranes shown in Fig. 4(F) and Fig. S10 (ESI†), where PhosWood-0 possesses a charge density of 0.1 mmol g⁻¹, while PhosWood-40 reaches 7.1 mmol g⁻¹. As illustrated in Fig. 4(G), the ionic conductivity of PhosWood-40 is an order of magnitude higher than that of nanosheet membranes, polymer hydrogels, nanocomposite materials and cellulose-based membranes.

Surface-charge-governed cation selectivity in phosphorylated bio-membranes

To further explore the ion transport mechanism in phosphorylated bio-membranes, we characterized the Zeta potential curves, energy dispersive spectroscopy (EDS) elemental spectra, and elemental content of different membranes. As illustrated in Fig. 5(A), the original zeta potential of unmodified cellulose is only −20.9 mV. Membranes with a higher degree of phosphorylation exhibit a higher negative zeta potential, with PhosWood-40 reaching −80.3 mV. This phenomenon aligns with previous findings, where phosphate groups dissociate into negatively charged ions on the cellulose surface, resulting in higher charge densities, and consequently, higher negative zeta potentials. To directly evaluate the cation selectivity ability, EDS elemental analysis (Fig. 5(B)–(D)) was performed to compare the levels of K, P, and Cl elements within the membranes. The content of P elements increases with higher degrees of phosphorylation, which is consistent with previous XPS results. Similarly, K element content inside the membranes exhibits an increasing trend as the increase of phosphorylation, indicating that phosphorylation enhances ion transport by expanding the effective ion channel space, as presented in Fig. S7B (ESI†). In addition, the K/Cl ratio gradually increases from 1.4 in PhosWood-0 to 54.3 in PhosWood-40 with the increase of P element content, reflecting the enhanced ionic selectivity of the membranes.

As negatively charged phosphate groups are introduced onto the surface of the cellulose molecular chain, the ion channels became sufficiently small (typically < 100 nm), which impedes the transport of anions, allowing more cations to pass through. This results in a cation-selective channel (Fig. 5(E)). To further understand the underlying mechanism of ionic conductivity (Fig. 5(F)–(H)), the state of water in phosphorylated bio-membranes was investigated using low-field nuclear magnetic resonance (LF-NMR). The T₂₁ peak corresponds to the minimum relaxation time range, indicating bound water with the lowest mobility, closely absorbed by hydrophilic groups on the molecular chain surface. The T₂₂ peak represents capillary water within the microporous structure, while the T₂₃ peak reflects free water in bulk, which has high mobility. In the confined swelling state (fully hydrated in water but confined by epoxy), only bound water is detected in PhosWood-40, whereas all three states of water are present in PhosWood-0. In the fully swollen state (freely soaked in water), Wood 40 manifests higher water absorption (22.3%) compared to PhosWood-0 (7.9%), though the proportion of capillary water in PhosWood-40 is only 0.46%. This suggests that most water molecules in PhosWood-40 are closely absorbed on the surface of cellulose molecular chains, facilitating the transport of mobile counterions. Fig. 5(I)–(K) depicts the mechanism of ionic conductivity across three types of membranes. PhosWood-0 exhibits low ionic conductivity due to the paucity of functional sites (Fig. 5(I)).

The state of water molecules is critical for ionic conductors;⁶¹ therefore, we analyzed the influence of water on ionic conductivity by comparing the effective ion channel space proportion and conductivity of phosphorylated bio-membranes under different swelling conditions. In Fig. 5(J), the mobility of counterion is improved as the introduction of electronegative phosphate groups, which increases both the effective ion channel space and conductivity in the confined swelling state (Fig. S7B and C, ESI†). Notably, the ionic conductivity of full swollen PhosWood-40 is higher than that in the confined swelling state at high KCl concentrations, but lower at low KCl concentrations (Fig. S7A, ESI†). This is because highly mobile water molecules facilitate ion diffusion in bulk solution, contributing to electrolyte-concentration-governed conductivity. In fully swollen PhosWood-40, the effective ion channel space proportion increases due to the presence of bulk free water. However, this free water causes the ionic conductivity in swollen PhosWood-40 to be more influenced by electrolyte concentrations within the channels, leading to reduced effective ion channel conductivity at low salt concentrations (Fig. 5(K) and Fig. S7C, ESI†). Hence, as water molecules adsorbed on the cellulose molecular chains promote ion translocation across phosphate groups, PhosWood-40 exhibits superior nanofluidic ion conductivity properties dominated by surface charge.

Salinity-gradient power generation performance of phosphorylated bio-membranes

To demonstrate the practical application potential of the phosphorylated bio-membranes, we evaluated their performance in salinity-gradient power harvesting. As mentioned earlier, these membranes exhibit favorable cation selectivity, preferentially transporting K⁺ from the high concentration side to the low concentration side, thereby generating diffusion currents and diffusion potentials (U). Fig. 6(A) displays the schematic setup of the electrochemical cell used for ionic current–voltage (I–V) measurements. In Fig. 6(B), the low concentration (LC) side contains a 10 mM L⁻¹ KCl solution, while the high concentration (HC) side contains a 500 mM L⁻¹ KCl solution. From the measured I–V curves obtained under transmembrane concentration gradients, the diffusion potential, and diffusion current density can be calculated after eliminating the redox part. The intersection of the I–V curve with the X- and Y-axes represents the diffusion voltage and current indensity of the membranes, respectively. Both values increase with higher phosphorylation levels due to enhanced cation selectivity, resulting in larger diffusion potentials in the nanofluidic membranes. For the practical assessment, we tested the membranes in a simulated seawater (0.5 M KCl) and river water (0.01 M KCl) system to power an external load (R_load). Fig. S14A and B (ESI†) depict the current density and power density of phosphorylated bio-membranes in this setup. As resistance increases, the diffusion current density (calculated as I = U/(R_load + R_membrane)) decreases, while the diffusion potential remains consistent. The output power density (calculated as I = U²/(R_load + R_membrane)) reaches a maximum at a resistance of about 6 kΩ. PhosWood-40 achieves a maximum output power density of up to 6.4 W m⁻², which is about 30 times that of PhosWood-0 (0.2 W m⁻²), highlighting its superior salinity-gradient power conversion performance of phosphorylated bio-membranes.

As presented in Fig. 6(D) and (E), maximum output power density, ion transference number (t₊) and energy conversion efficiencies (η) are proportional to the phosphorylation degrees of PhosWood. Since the ion transference number (t₊) generally represents the cation selectivity of the membrane,⁶² the strong cation selectivity of PhosWood, which is close to the ideal cation value of 1, contributes to the higher diffusion voltage and energy conversion efficiencies. Considering practical operating conditions, we further investigated the performance of PhosWood-40 across different concentration gradients, with the LC side fixed at 10 mM L⁻¹ KCl. The concentration gradient is regulated by the KCl solution in the HC side. At a concentration gradient of 1 (equal salt concentrations on both sides), neither diffusion potential nor current could be detected, as there is no directional ion diffusion. With the increase of KCl concentration, the directional diffusion of K⁺ within the membrane is enhanced, leading to higher diffusion current densities and diffusion voltages. The maximum output power density of PhosWood-40 increases from 1.2 to 11.9 W m⁻² with an increasing KCl concentration gradient from 5 to 300-fold (Fig. 6(F)). As the concentration of KCl increases, the reduction in Debye length under high-concentration conditions leads to a decrease in ion transference number and energy conversion efficiencies. Notably, PhosWood-40 maintained excellent performance in all ranges of salt concentration gradient environments with a t₊ from 0.98 to 0.92 and η from 49% to 35%, which may meet the requirements for applications across a range of salinity-gradient scenarios.

Additionally, membranes with higher phosphorylation exhibit increased ionic conductivity (Fig. 6(G)), which reduces transmembrane resistance and further increases current intensity. This result is consistent with findings on surface-charge-governed cation selectivity. Consequently, the superior osmotic energy conversion ability can be attributed to the increase of the modified surface negative charge density and the reserved natural highly aligned structure within PhosWood-40. Theoretically, thinner membranes generate higher power densities due to faster ion diffusion. In practice, however, block materials with highly ordered internal structures are difficult to process into thin films. Even though thinner ordered membranes can be prepared using nanosheet membranes, polymer, and nanocomposites to achieve good performance, the ion channels in these materials are oriented perpendicular to the desired direction of ion transport. In comparison with other reported nanofluidic materials, our phosphorylated bio-membrane demonstrates significant advantages at the same thickness (Fig. 6(H)).

Conclusions

In conclusion, we have developed a phosphorylation-assisted cell wall engineering strategy to fabricate highly ion-conductive and mechanically robust bio-membranes from natural wood, achieving an optimal balance between high negative charge density and multiscale structural alignment. By controlling the introduction of phosphate groups by a phosphorylation process, we significantly enhanced surface charge density while preserving the hierarchical alignment of natural pinewood from macroscopic to molecular scales, resulting in enhanced strength and ionic conductivity. Notably, a 40-minitue phosphorylation process yielded the highest charge density of 7.1 mmol g⁻¹, which dramatically increased surface electronegativity. This enhancement led to a remarkable ionic conductivity of 21.01 mS cm⁻¹ in 1.0 × 10⁻⁵ mol L⁻¹ KCl aqueous solution, representing a 100-fold increase compared to unmodified membranes, along with significantly improved cation selectivity. The ion conductivity and salinity-gradient power generation performance of these phosphorylated bio-membranes far surpass those of previously reported nanofluidic membranes, including cellulose-based materials, polymers, and composite nanomaterials. This work provides novel insights into the design and fabrication of cellulose-based nanofluidic membranes and opens new avenues for their application in nanofluidic energy capture, sensors, energy storage, and advanced ion-conducting devices.

Material and methods

Materials and chemicals

Pinewood (80 mm × 80 mm × 1 mm) was provided by Guiyang Wood Industry Co., Ltd (Guiyang, China). Sodium hydroxide (AR, 95%), sodium chloride (AR, 99.5%), potassium chloride (AR, 99.5%), urea (AR, 99%), diammonium hydrogen phosphate (AR, 99.5%), were purchased from Macklin Inc. Hydrochloric acid (37%), sodium sulfite (>97%), methanol (>99.5%), ethanol (absolute) and anthraquinone (>98%) were purchased from Guangzhou Chemical Reagent Co., Ltd (China).

Preparation of phosphorylated bio-membranes

A phosphorylation-assisted cell wall engineering strategy was used to fabricate ultra-strong and high ion-conductive bio-membranes from natural trees. The entire preparation process is illustrated in Fig. S12 (ESI†). Delignified pinewood was prepared according to our previous work.⁶³ The delignified samples were initially mixed with (NH₄)₂HPO₄, and urea in a molar ratio of 1:1.5:10, followed by drying at 70 °C for 12 hours to ensure complete water removal. The phosphorylation reaction was carried out at 170 °C for varying durations (5, 10, 15, 20, 40, and 60 minutes) under controlled conditions to achieve different degrees of phosphate group incorporation. After heating, the samples were washed with deionized water until the conductivity of the filtrate was below 10 μS cm⁻¹.

Subsequently, the phosphorylated samples were compressed into bio-membranes using a hot press at a maximum pressure of 10 MPa and a temperature of 90 °C. Delignified wood membranes were labeled as PhosWood-0, while those subjected to phosphorylation for 5, 10, 15, 20, 40, and 60 minutes were labeled as PhosWood-5, PhosWood-10, PhosWood-15, PhosWood-20, PhosWood-40, and PhosWood-60, respectively. The lignin, cellulose, and hemicellulose content of natural wood, delignified wood and PhosWood-40 are shown in Fig. S13 (ESI†).

Analysis and testing procedures

The FI-IR spectra of membranes were measured using a FT-IR spectrometer (VERTEX 70, Germany, Bruker). Elemental analysis of the membranes was characterized by X-ray photoelectron spectroscopy (XPS) (250Xi, Thermo Fisher, American). X-ray diffraction (XRD) was performed by Bruker D8 instrument (Bruker, Germany) at a scanning speed of 12° min⁻¹ from 5° to 40° (40 mA, 40 kV). Water absorption was determined according to ASTM D570-98.⁶⁴ The Negative charge density of cellulose was performed following a previously reported method.⁶⁵ The morphologies of membranes were observed using an optical microscope (BX51, Olympus, Japan) and a field-emission scanning electron microscope (SEM, HITACHI, Japan). Small-angle X-ray scattering (SAXS) measurements of membranes were performed using a SAXS system (Xenocs, France). General tensile tests were carried out on a material experiment machine (INSTRON 5565, America) at 40 RH% humidity and 25 °C, the dry samples are placed in the condition of 40 RH% humidity and 25 °C for 24 hours before the text. The tensile strength of wet samples is texted according to standard GB/T 465.2-2008. For ion conductivity measurements, electrochemical impedance spectroscopy (EIS) and current–voltage (I–V) curve were conducted using an electrochemical workstation (CHI660E, China). To prepare the samples, 5 mm wide membranes were secured in epoxy resin, with reservoirs carved at both ends to facilitate electrode contact. The epoxy resin helped maintain the structural integrity of the membrane in water. Before conductivity measurements, the nanofluid device was immersed in KCl solution with concentrations ranging from 10⁻⁶–1 M for 48 hours. Ionic conductivity (κ, mS cm⁻¹) was calculated based on the length (l, cm), width (w, cm), height (h, am) and resistance (R, Ω) of the swollen membranes in epoxy resin: κ = l/Rwh. Zeta potential analysis was performed using a zeta Potential Analyzer (zeta Plus, Brookhaven Instruments Corporation, USA) at 25 °C, after dispersing the samples in deionized water to create 0.1 wt% dispersions using a microfluidizer processor (M110EH, Microfluidics Ind., USA). The low-field nuclear magnetic resonance transverse relaxation time (LF-NMR (T2)) experiments were measured by medium-sized NMR analysis and imaging system (MesoMR23-060H-I, Suzhou Niumag, China).

Author contributions

K. Chen and Z. Fang conceived and designed the experiments. K. Chen, J. Zhou, and L. Yu conducted the experiments. K. Chen, Z. Fang, and C. Chen analysed the data and created schematics. K. Chen and Z. Fang drafted the manuscript. X. Qiu oversaw project administration, supervision and conceptualization. C. Chen, C. Xu and J. Zhou contributed to review & editing. All authors commented on the final manuscript.

Conflicts of interest

There are no conflicts to declare.

Data availability

The authors confirm that the data supporting the findings of this study are available within the article [and/or] its ESI.†

Acknowledgements

This work was supported by the Fund of the National Natural Science Foundation of China (U23A6005), Guangdong Basic and Applied Basic Research Foundation (2025A1515010005, 2023B1515040013, 2020B1515020021), and the State Key Laboratory of Pulp & Paper Engineering (2024C03, 2023C05, 2022C01).

References

1. Y. Zhou, X. Liao, J. Han, T. Chen and C. Wang, Chin. Chem. Lett., 2020, 05, 33 Search PubMed .
2. L. Li, X. Sun, J. Miao, H. Wang, Y. Song and D. Tang, Energy Environ. Sci., 2025, 18, 1176–1190 RSC .
3. C. Yang and Z. Suo, Nat. Rev. Mater., 2018, 3, 125–142 CrossRef CAS .
4. W. He, L. Xu, G. Yu, K. Wang, D. Bao, X. Tang, X. Zhou, J. Zhang, T. Huang, N. Li, M. Guo, Y. Yue and Y. Ma, Adv. Mater., 2025, 37, 2417498 CrossRef CAS PubMed .
5. Z. Yang, H. Liu, J. Li, K. Yang, Z. Zhang, F. Chen and B. Wang, ACS Appl. Mater. Interfaces, 2020, 12, 15002–15011 CrossRef CAS PubMed .
6. S. Li, Y. Li, Y. Shao and H. Wang, Adv. Fiber Mater., 2021, 4, 129–144 CrossRef .
7. L. Wang, S. Xie, Z. Wang, F. Liu, Y. Yang, C. Tang, X. Wu, P. Liu, Y. Li, H. Saiyin, S. Zheng, X. Sun, F. Xu, H. Yu and H. Peng, Nat. Biomed. Eng., 2020, 4, 159–171 CrossRef CAS PubMed .
8. J. Gao, Y. Feng, W. Guo and L. Jiang, Chem. Soc. Rev., 2017, 46, 5400–5424 RSC .
9. C. Zhu, R. E. Usiskin, Y. Yu and J. Maier, Science, 2017, 358, eaao2808 CrossRef PubMed .
10. S. Li, Z. Fan, G. Wu, Y. Shao, Z. Xia, C. Wei, F. Shen, X. Tong, J. Yu, K. Chen, M. Wang, Y. Zhao, Z. Luo, M. Jian, J. Sun, R. B. Kaner and Y. Shao, ACS Nano, 2021, 15, 7821–7832 CrossRef CAS .
11. A. van den Berg, H. G. Craighead and P. Yang, Chem. Soc. Rev., 2010, 39, 899 RSC .
12. W. Sparreboom, A. van den Berg and J. C. T. Eijkel, Nat. Nanotechnol., 2009, 4, 713–720 CrossRef CAS PubMed .
13. D. Stein, M. Kruithof and C. Dekker, Phys. Rev. Lett., 2004, 93, 35901 CrossRef PubMed .
14. R. B. Schoch and P. Renaud, Appl. Phys. Lett., 2005, 86, 253111 CrossRef .
15. K. Xiao, L. Jiang and M. Antonietti, Joule, 2019, 3, 2364–2380 CrossRef CAS .
16. P. Wang, W. Tao, T. Zhou, J. Wang, C. Zhao, G. Zhou and Y. Yamauchi, Adv. Mater., 2024, 36, 2404418 CrossRef CAS PubMed .
17. M. Zhang, R. Guo, K. Chen, Y. Wang, J. Niu, Y. Guo, Y. Zhang, Z. Yin, K. Xia, B. Zhou, H. Wang, W. He, J. Liu, M. Sitti and Y. Zhang, Proc. Natl. Acad. Sci. U. S. A., 2020, 117, 14667–14675 CrossRef PubMed .
18. S. Hong, C. Constans, M. V. Surmani Martins, Y. C. Seow, J. A. Guevara Carrió and S. Garaj, Nano Lett., 2017, 17, 728–732 CrossRef CAS PubMed .
19. K. Raidongia and J. Huang, J. Am. Chem. Soc., 2012, 134, 16528–16531 CrossRef CAS PubMed .
20. K. Xiao, P. Giusto, L. Wen, L. Jiang and M. Antonietti, Angew. Chem., Int. Ed., 2018, 57, 10123–10126 CrossRef CAS PubMed .
21. S. Qin, D. Liu, Y. Chen, C. Chen, G. Wang, J. Wang, J. M. Razal and W. Lei, Nano Energy, 2018, 47, 368–373 CrossRef CAS .
22. M. Liu, M. Huang, L. Tian, L. Zhao, B. Ding, D. Kong, Q. Yang and J. Shao, ACS Appl. Mater. Interfaces, 2018, 10, 44915–44923 CrossRef CAS PubMed .
23. J. J. Shao, K. Raidongia, A. R. Koltonow and J. Huang, Nat. Commun., 2015, 6, 7602 CrossRef PubMed .
24. C. Zhu, P. Liu, B. Niu, Y. Liu, W. Xin, W. Chen, X. Kong, Z. Zhang, L. Jiang and L. Wen, J. Am. Chem. Soc., 2021, 143, 1932–1940 CrossRef CAS PubMed .
25. X. Wang, Z. Wang, Z. Xue, Y. Fan, J. Yang, Q. Zhang, N. Yang, X. Meng, Y. Jin and S. Liu, Energy Environ. Sci., 2024, 17, 6225–6233 RSC .
26. R. K. Gogoi, A. B. Neog, T. J. Konch, N. Sarmah and K. Raidongia, J. Mater. Chem. A, 2019, 7, 10552–10560 RSC .
27. J. Lao, R. Lv, J. Gao, A. Wang, J. Wu and J. Luo, ACS Nano, 2018, 12, 12464–12471 CrossRef CAS PubMed .
28. S. Zhou, Y. Hu, W. Xin, L. Fu, X. Lin, L. Yang, S. Hou, X. Y. Kong, L. Jiang and L. Wen, Adv. Mater., 2023, 35, 2208640 CrossRef CAS .
29. L. Ding, D. Xiao, Z. Zhao, Y. Wei, J. Xue and H. Wang, Adv. Sci., 2022, 9, e2202869 CrossRef PubMed .
30. J. Li, J. P. Singh, V. Neklyudov, M. Stolov, Z. Yuan, Y. Schilt, U. Raviv, D. R. Dekel and V. Freger, Sci. Adv., 2024, 10, eadp1450 CrossRef CAS PubMed .
31. C. Montanari, Y. Li, H. Chen, M. Yan and L. A. Berglund, ACS Appl. Mater. Interfaces, 2019, 11, 20465–20472 CrossRef CAS PubMed .
32. A. Awati, R. Yang, T. Shi, S. Zhou, X. Zhang, H. Zeng, Y. Lv, K. Liang, L. Xie, D. Zhu, M. Liu and B. Kong, Angew. Chem., Int. Ed., 2024, 63, e202407491 CrossRef CAS PubMed .
33. X. Zhao, C. Lu, L. Yang, W. Chen, W. Xin, X. Kong, Q. Fu, L. Wen, G. Qiao and L. Jiang, Nano Energy, 2021, 81, 105657 CrossRef CAS .
34. Y. Jiang, R. Hu, C. Yang, Z. Zhou, G. Yuan, H. Zhou and S. Hu, Sci. Adv., 2023, 9, 8493 CrossRef PubMed .
35. J. Ji, Q. Kang, Y. Zhou, Y. Feng, X. Chen, J. Yuan, W. Guo, Y. Wei and L. Jiang, Adv. Funct. Mater., 2017, 27, 1603623 CrossRef .
36. X. Li, X. Qiu, X. Yang, P. Zhou, Q. Guo and X. Zhang, Adv. Mater., 2024, 2407170 CrossRef CAS PubMed .
37. C. Chen and L. Hu, Adv. Mater., 2020, 2002890 Search PubMed .
38. W. Kong, T. Li, C. Chen, G. Chen, A. H. Brozena, D. Liu, Y. Liu, C. Wang, W. Gan, S. Wang, S. He and L. Hu, Chem. Mater., 2019, 31, 9288–9294 CrossRef CAS .
39. N. Mittal, F. Ansari, V. K. Gowda, C. Brouzet, P. Chen, P. T. Larsson, S. V. Roth, F. Lundell, L. Wågberg, N. A. Kotov and L. D. Söderberg, ACS Nano, 2018, 12, 6378–6388 CrossRef CAS PubMed .
40. W. Xin, Z. Zhang, X. Huang, Y. Hu, T. Zhou, C. Zhu, X. Kong, L. Jiang and L. Wen, Nat. Commun., 2019, 10, 3876 CrossRef PubMed .
41. C. Wang, C. Miao, X. Zhu, X. Feng, C. Hu, D. Wei, Y. Zhu, C. Kan, D. Shi and S. Chen, ACS Appl. Nano Mater., 2019, 2, 4193–4202 CrossRef CAS .
42. Y. Zhou, C. Chen, X. Zhang, D. Liu, L. Xu, J. Dai, S. Liou, Y. Wang, C. Li, H. Xie, Q. Wu, B. Foster, T. Li, R. M. Briber and L. Hu, J. Am. Chem. Soc., 2019, 141, 17830–17837 CrossRef CAS PubMed .
43. N. Sheng, S. Chen, M. Zhang, Z. Wu, Q. Liang, P. Ji and H. Wang, ACS Appl. Mater. Interfaces, 2021, 13, 22416–22425 CrossRef CAS .
44. L. A. Berglund and I. Burgert, Adv. Mater., 2018, 30, 1704285 CrossRef PubMed .
45. J. Huang, B. Zhao, T. Liu, J. Mou, Z. Jiang, J. Liu, H. Li and M. Liu, Adv. Funct. Mater., 2019, 29, 1902255 CrossRef .
46. C. Chen, Y. Kuang, S. Zhu, I. Burgert, T. Keplinger, A. Gong, T. Li, L. Berglund, S. J. Eichhorn and L. Hu, Nat. Rev. Mater., 2020, 5, 642–666 CrossRef CAS .
47. T. Li, S. X. Li, W. Kong, C. Chen, E. Hitz, C. Jia, J. Dai, X. Zhang, R. Briber, Z. Siwy, M. Reed and L. Hu, Sci. Adv., 2019, 5, eaau4238 CrossRef CAS .
48. T. Li, X. Zhang, S. D. Lacey, R. Mi, X. Zhao, F. Jiang, J. Song, Z. Liu, G. Chen, J. Dai, Y. Yao, S. Das, R. Yang, R. M. Briber and L. Hu, Nat. Mater., 2019, 18, 608–613 CrossRef CAS PubMed .
49. Q. Y. Wu, C. Wang, R. Wang, C. Chen, J. Gao, J. Dai, D. Liu, Z. Lin and L. Hu, Adv. Energy Mater., 2020, 10, 1902590 CrossRef CAS .
50. W. Kong, C. Chen, G. Chen, C. Wang, D. Liu, S. Das, G. Chen, T. Li, J. Li, Y. Liu, Z. Li, B. C. Clifford and L. Hu, Small, 2021, 17, 2008200 CrossRef CAS PubMed .
51. G. Chen, T. Li, C. Chen, C. Wang, Y. Liu, W. Kong, D. Liu, B. Jiang, S. He, Y. Kuang and L. Hu, Adv. Funct. Mater., 2019, 29, 1902772 CrossRef CAS .
52. G. Chen, T. Li, C. Chen, W. Kong, M. Jiao, B. Jiang, Q. Xia, Z. Liang, Y. Liu, S. He and L. Hu, ACS Nano, 2021, 15, 11244–11252 CrossRef CAS PubMed .
53. B. Thomas, M. C. Raj, A. K. B, R. M. H, J. Joy, A. Moores, G. L. Drisko and C. Sanchez, Chem. Rev., 2018, 118(24), 11575–11625 CrossRef CAS PubMed .
54. P. Zhou, G. Jiang, Y. Wang, Y. Tian and X. Zhang, Nat. Commun., 2025, 16, 3791 CrossRef CAS PubMed .
55. C. Zhou, Z. Chen, H. Yang, K. Hou, X. Zeng, Y. Zheng and J. Cheng, ACS Appl. Mater. Interfaces, 2017, 9, 9184–9194 CrossRef CAS PubMed .
56. Y. Matsuo, T. Kimura, H. Koyanagi, N. Netsu, F. Komatsu, T. Nagata, S. Nishitsuji, J. Matsui and A. Masuhara, Green Chem., 2025, 27, 3532–3541 RSC .
57. N. Kumari, M. Jassal and A. K. Agrawal, Carbohydr. Polym., 2021, 256, 117531 CrossRef CAS PubMed .
58. F. Rol, C. Sillard, M. Bardet, J. R. Yarava, L. Emsley, C. Gablin, D. Léonard, N. Belgacem and J. Bras, Carbohydr. Polym., 2020, 229, 115294 CrossRef CAS PubMed .
59. Y. Habibi, Chem. Soc. Rev., 2014, 43, 1519–1542 RSC .
60. A. Paajanen, A. Zitting, L. Rautkari, J. A. Ketoja and P. A. Penttilä, Nano Lett., 2022, 22, 5143–5150 CrossRef CAS PubMed .
61. T. Emmerich, N. Ronceray, K. V. Agrawal, S. Garaj, M. Kumar, A. Noy and A. Radenovic, Nat. Rev. Methods Primers., 2024, 4, 69 CrossRef CAS .
62. W. Jiang, J. Zhou, X. Zhong, M. Fang, J. Hao, D. Zhao, X. Wen, H. Wang, Y. Zhou, Y. Zhu and L. Jiang, Nat. Sustain., 2025, 8, 446–455 CrossRef .
63. Z. Fang, B. Li, Y. Liu, J. Zhu, G. Li, G. Hou, J. Zhou and X. Qiu, Matter, 2020, 2, 1000–1014 CrossRef .
64. Astm. Standard Test Method for Water Absorption of Plastics (ASTM D570-98), 10.1520/D0570-98R18.
65. K. Chen, L. Peng, Z. Fang, X. Lin, C. Sun and X. Qiu, Carbohydr. Polym., 2023, 121250 CrossRef PubMed .

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d5mh01003a

‡ Equal contribution to this manuscript.

This journal is © The Royal Society of Chemistry 2025