Synchronously reconfiguring closed pore and interlayer spacing of wood-derived hard carbon via hot-pressing for advanced sodium-ion batteries

Abstract

Biomass-derived hard-carbon materials have attracted considerable interest due to their abundant availability and high sodium storage capacities. However, optimizing the sodium storage performance of these materials requires precise control of the components and microstructure of the biomass. Herein, we report a simple and effective hot-press densification strategy to optimize the closed-pore structure and microcrystalline properties of carbonized wood fiber (CWF), significantly improving the platform capacity and initial coulombic efficiency (ICE) of sodium-ion batteries (SIBs). These results demonstrate that an appropriate hot-pressing treatment promotes the recrystallization of amorphous cellulose in WF, facilitates the decomposition of hemicellulose, and contributes to the formation of more closed pores, larger interlayer spacing, and smaller specific surface area during high-temperature carbonization. Molecular dynamics simulations are employed to elucidate the mechanism by which an increase in defect structures during hot pressing leads to an expansion of the interlayer spacing and an increase in sp² carbon content during pyrolysis. Under the conditions of an initial moisture content of 50% of WF and a pressing pressure of 6 MPa, the densified carbonized wood fibers (DCWF-6) exhibit a high specific capacity of 427.1 mA h g⁻¹ at a current density of 0.1 A g⁻¹, with an ICE of 86%. In addition, they show excellent rate performance, maintaining a specific capacity of 197.7 mA h g⁻¹ at a current density of 4.0 A g⁻¹. This simple, low-cost hot-press densification strategy is highly effective and holds great promise for improving the energy density of SIBs with potential applicability to other biomass precursors.

---

Green foundation

1. This work advances green chemistry by developing a low-cost, energy-efficient method to enhance sodium-ion battery performance using biomass-derived hard-carbon materials. It optimizes the structure of carbonized wood fibers through hot-press densification, promoting sustainability in energy storage.

2. The key achievement lies in using a green and efficient hot-press densification technique under moist heat conditions to promote the recrystallization of amorphous cellulose, thereby enhancing the sodium storage capacity and initial coulombic efficiency of wood fiber-derived hard carbon. This approach offers advantages over previously reported composition-controlled methods.

3. Future work could enhance sustainability by exploring energy-efficient densification techniques and using renewable energy in processing. Optimizing biomass feedstocks, such as waste materials, and scaling the process can further reduce environmental impact and broaden its applicability in energy storage.

1. Introduction

Since the release of the first-generation commercial sodium-ion batteries (SIBs) by CATL, a new energy storage journey has begun. SIBs have garnered significant attention as rechargeable battery technologies due to the abundant sodium resources, low cost, and high safety performance. They have vast application prospects in energy storage, electric vehicles, and renewable energy.^1–3 Compared with the rapidly developing SIB cathode materials, the development of high-performance anode materials is relatively lacking, which has become a significant factor restricting the power density, energy density, and rapid commercialization of SIBs.^4–6 Therefore, innovative, affordable, feasible, high-performance anode materials are required.

Among the many SIB anode materials, hard carbon is currently an ideal candidate due to its low cost, low working potential (∼0.2 V), and good cycling life. Hard-carbon materials have complex disordered structures composed of randomly dispersed, bent, and defective graphite-like layers.⁷ As an anode material for SIBs, the charge–discharge curve of hard carbon mainly consists of sloping and plateau regions. Existing research indicates that the capacity in the sloping region is related to defects and surface functional groups of the carbon layers, whereas the capacity in the plateau region is associated with a closed pore structure within the carbon layers. Simultaneously, increasing the plateau-region capacity of hard-carbon anodes can effectively enhance the overall capacity and energy density of SIBs.^8–10 Therefore, a deep understanding of the sodium storage mechanism and the construction of more closed pore structures in hard-carbon materials is the key to accelerating the development of SIBs.

The properties of the hard-carbon precursors significantly affect the structural composition of the derived hard carbon materials. Biomass is considered the most promising precursor of hard carbon due to its abundance, low cost, renewability, and environmental friendliness.^11,12 Plants, as the most abundant natural resource on Earth, have hierarchical micro/nanostructures, rich hydroxyl functional groups, and high carbon content, making them ideal hard-carbon precursors.^13,14 Tang et al.¹ used waste wood as a carbon source and prepared hard-carbon anode materials with a highly closed pore structure by regulating its components. Research has indicated that crystalline and amorphous celluloses are key factors in achieving closed-pore structures during pyrolysis. High-content crystalline cellulose can generate more cross-linked disordered graphite-like layers, which further wind and fold to form walls with closed-pore structures. Subsequently, Wang et al.¹⁵ used bamboo as a hard-carbon precursor to optimize the closed-pore structure of a hard-carbon material by regulating its lignin content. This study found that removing part of the lignin from the bamboo-exposed radicals in cellulose and hemicellulose resulted in the production of more small molecules and more intense radical reactions during pyrolysis. However, the excessive removal of lignin can limit the formation of closed-pore structures. Recently, Zhang et al.¹⁶ used lignocellulose as a carbon precursor to optimize the structural composition of hard-carbon materials by regulating their components. This study found that lignocellulose with a low hemicellulose content can be used to prepare hard-carbon materials with higher capacities. The aforementioned research indicates that by regulating the composition of biomass precursors, the closed-pore structure of hard-carbon materials can be effectively optimized, thereby increasing their specific capacity. However, these processes are relatively complex, and the initial coulombic efficiency (ICE) of the resulting hard-carbon anodes requires further improvement.

Wood fiber (WF), which consists of a three-dimensional network of natural cellulose I nanofibers, is one of the most abundant and sustainable biopolymers on Earth. These nanofibers have a high Young's modulus and numerous hydroxyl groups that can serve as active sites for carbon materials. WF is derived from wood processing residues, making it an environmentally sustainable, abundant, and renewable resource, as well as a valuable fibrous material for the paper industry. However, WF has yet to be utilized in a valuable manner.^17,18

To address these issues, this study employed a simple hot-press densification pretreatment to regulate the structural composition of carbonized WF (CWF) and reduce its specific surface area, thereby adjusting its closed-pore structure. During the hot-press densification process, some hemicellulose was removed, enhancing the cross-linking between lignin and promoting the conversion of amorphous cellulose to crystalline cellulose. The partial removal of cellulose and cross-linking between lignins helps inhibit graphitization and promote the formation of closed pores, whereas an increase in crystalline cellulose regulates the wall layers of the closed-pore structure. By contrast, excessive pressure during densification can damage the crystalline cellulose, reducing the formation of closed pores. The emergence of numerous closed pore structures results in a CWF with a higher plateau capacity. Moreover, the densification pretreatment reduced the specific surface area of the CWF, decreasing the solid–liquid interfacial reactions and increasing the ICE. This study is significant for expanding the high-value utilization of WF and advancing the development of efficient and economical hard-carbon anode materials.

2. Experimental section

2.1 Materials

The water used in this study was ultrapure water, and the natural wood fibers used were derived from poplar.

2.2 Synthesis of DCWF-P

50 g of WF (moisture content of 50%) on a hot press machine and adjust the hot press pressure. The hot press time is 10 min, and the hot press temperature is 150 °C. After hot pressing is completed, remove the sample and label it as DWF-P (P represents the processing pressure). Subsequently, place DWF-P in a tube furnace and carbonize it at 1300 °C for 2 hours under a nitrogen atmosphere with a heating rate of 2 °C min⁻¹. The resulting sample is labeled as DCWF-P. Using a preparation method similar to that of DCWF-P, products with varying initial moisture contents (denoted as x, where x = 20% and 80%) were obtained, resulting in different products (DCWF-20% and DCWF-80%). Additionally, as a comparison, directly carbonize the WF without hot pressing to obtain the product labeled CWF.

3. Results and discussion

3.1 Preparation and characterizations

The preparation of DCWF-P via hot-press densification is illustrated in Fig. 1. The WF underwent hot-press treatment to obtain densified WFs (DWF-P), which were further carbonized at high temperatures to yield densified CWFs (DCWF-P) (Fig. 1a). After the hot-press densification treatment of the WF, the amorphous cellulose underwent recrystallization, partial decomposition of hemicellulose, and condensation of lignin (Fig. 1b). During carbonization, crystalline cellulose can form cross-linked, disordered, and graphite-like layers. These graphite-like layers undergo shrinkage and entanglement during the subsequent high-temperature carbonization, forming pore walls for a closed-pore structure. In addition, the transformations in hemicellulose and lignin accelerate the pyrolysis of the WF, facilitating the formation of abundant active sites during the low-temperature pyrolysis of the DWF. These active sites act as anchors, preventing the melting and ordered rearrangement of crystalline cellulose during high-temperature carbonization. This inhibits excessive graphitization of the hard-carbon material, promotes the formation of a closed-pore structure, and increases the interlayer spacing within the carbon matrix (Fig. 1c).

Fig. 1 Schematic illustration of DCWF-P hard carbon material synthesized by wood fiber (a). Densification-assisted regulation of composition (b) and its impact on closed-pore formation mechanism (c).

To investigate the effects of hot-press densification treatment on the crystallinity of WF, X-ray diffraction (XRD) analysis was conducted on DWF-P and DWF-x samples. Fig. 2a and S1a† show the XRD spectra of WF, DWF-P and DWF-x. All samples exhibit characteristic peaks of cellulose (101) and (200) at 2θ = 15.5° and 22.5°.^19,20 Compared to WF, the intensities of the cellulose peaks in DWF-4 and DWF-6 were enhanced, indicating that hot-press densification treatment causes molecular chain breakage in the non-crystalline regions of cellulose, gradually reducing the amorphous region. This exposure of hydroxyl groups on the surface of the microfibrils in both the crystalline and amorphous regions promotes the formation of van der Waals forces and hydrogen bonds between adjacent cellulose chains, thereby enhancing the internal structural stability of the WF.^21,22 However, when the pressure increased to 8 MPa, the intensity of the cellulose peak in DWF-8 decreased. This is attributed to the excessive pressure restricting the interactions between cellulose molecules, thereby reducing cellulose crystallinity. In addition, as the initial moisture content in the WF increased, the intensity of the characteristic cellulose peaks gradually increased. This suggests that, although a hydrothermal environment may cause the partial degradation of hemicellulose, it exposes more hydroxyl groups, thereby increasing the hydrogen bonding area between the WF.²³ Crystallinity calculations revealed that WF, DWF-4, DWF-6, DWF-8, DWF-20% and DWF-80% had crystalline cellulose contents of 58.08%, 63.24%, 65.79%, 64.01%, 62.01%, and 67.11%, respectively. Fig. 2b and S1b† show the FTIR spectra of WF, DWF-P, and DWF-x. The spectra clearly show that no significant structural changes occurred in any of the samples, and identical characteristic peaks were observed. Notably, increasing the processing pressure and initial moisture content of WF led to a stronger and broader hydroxyl absorption peak at 3344 cm⁻¹.^24,25 This effect was primarily due to the enhanced bonding strength between WF under pressure, as well as the partial degradation and dissolution of hemicellulose in the hydrothermal environment, which exposed more hydroxyl groups on the WF cell walls, thereby promoting hydrogen bonding between WF.²⁶ These results are consistent with those of the XRD analysis. Fig. S1c† shows the thermogravimetric differential curves for WF and DWF-P. The mass loss in the range of 200 °C to 250 °C is primarily attributed to the decomposition of hemicellulose. Mass loss between 250 °C to 350 °C is mainly due to cellulose decomposition, whereas the slow mass loss from 350 °C to 800 °C derives from the further decomposition of C–C and C–H bonds, leading to the continued degradation of WF. In addition, the slow decomposition rate of lignin contributes to the gradual thermal degradation of the WF.²⁷ According to the TGA curves of the WF and DWF-P (Fig. S1d†), the mass losses of WF, DWF-4, DWF-6, and DWF-8 are 85.51%, 82.06%, 80.63%, and 80.31%, respectively. As the pressure increased, the mass loss of the WF gradually decreased, indicating that the hot-press densification treatment led to the partial decomposition of hemicellulose. In addition, the Van Soest method was used to quantitatively analyze the cellulose, hemicellulose, and lignin contents of WF and DWF-6. As Fig. 2c shows, compared to WF, DWF-6 exhibits an increase in the percentage of cellulose, a decrease in hemicellulose, and a slight decrease in the percentage of lignin. These results further indicate that the hot-press densification treatment leads to the partial decomposition of hemicellulose.

Fig. 2 Morphology and architecture of DCWF-P. XRD patterns (a) and Raman spectra (b) of WF and DWF-P. The percentage content (c) of cellulose, hemicellulose, and lignin in WF and DWF-6. SEM images of DWF-4 (d), DWF-6 (e), DWF-8 (f), DCWF-4 (g), DCWF-6 (h) and DCWF-8 (i). HRTEM images of DCWF-4 (j), DCWF-6 (k) and DCWF-8 (l).

To reveal the impact of the hot-press densification treatment on the morphology of the WF-derived hard-carbon materials, scanning electron microscopy (SEM) analysis was conducted on the prepared samples. Fig. S2a† and Fig. 2d–f show the SEM images of WF and DWF-P. As observed, the WF was randomly intertwined, and its surface was rough. Following the hot-press densification treatment, DWF-4 exhibited slight aggregation. As the pressure increased, the fibers flattened and the gaps between the fibers decreased. When the pressure reached 8 MPa, the surface of DWF-8 became dense and smooth, and the fibers were tightly intertwined. As Fig. S2b, S3,† and Fig. 2g–i show, after high-temperature carbonization, the corresponding CWFs shrunk and retained their natural one-dimensional hollow structure, which is beneficial for reducing electrolyte decomposition and the formation of the solid electrolyte interphase (SEI). To further investigate the effects of hot-press densification on the carbon layers and pore structure of the CWFs, high-resolution TEM (HRTEM) analysis was conducted. Fig. S2c† and Fig. 2j–l present HRTEM images of CWF and DCWF-P. The carbon layer structure of the CWF was amorphous, and the arrangement of the carbon layers exhibited anisotropy with no obvious long graphitic layers or closed pore structures. Following the hot-press densification treatment of WF, DCWF-4 showed longer graphitic layers (marked by white rectangles) and few closed pore structures (marked by white circles). As the pressure increased, the closed-pore structures in DCWF-6 became more prevalent. Notably, when the pressure reached 8 MPa, the closed pore structures in DCWF-8 decreased, and the graphitic layers also became shorter, which may have been due to the excessive pressure damaging the WF structure. Previous studies have shown that an increase in closed pore structures and the elongation of graphitic layers in hard carbon materials may be related to the increased content of crystalline cellulose in the precursor.²⁸ During the high-temperature carbonization process, the pyrolysis of crystalline cellulose forms longer graphitic layers, promoting the distortion, folding, and shrinkage of the carbon layers, thereby forming closed pore structures.

XRD and Raman spectroscopy were used to investigate the crystal structure and defect characteristics of CWF, DCWF-P, and DCWF-x. Fig. 3a and S4a† show the XRD patterns of CWF, DCWF-P and DCWF-x. All samples exhibit a typical amorphous carbon (0 0 2) diffraction peak at 2θ = 24.1°.²⁹ Compared to CWF, the (0 0 2) diffraction peak intensities of DCWF-4, DCWF-6, DCWF-20%, and DCWF-80% were enhanced, indicating that moderate pressurization and increased initial moisture content in WF are conducive to improving the graphitization degree of hard-carbon materials. However, when the treatment pressure was excessively high, a slight decrease in the (0 0 2) diffraction peak intensity of DCWF-8 was observed, suggesting that excessive pressure may disrupt the structure of the WF, resulting in reduced conductivity of DCWF-8 and slowing of ion/electron transport. In addition, a negative shift in the (0 0 2) diffraction peak is observed for DCWF-4, DCWF-6, DCWF-20%, and DCWF-80%. According to Bragg's law,³⁰ the interlayer spacing of carbon in the CWF increased from 0.375 to 0.386 nm (DCWF-4), 0.395 nm (DCWF-6), 0.382 nm (DCWF-8), 0.380 nm (DCWF-20%), and 0.402 nm (DCWF-80%). Moreover, the Lc value decreases from 1.01 nm in CWF to 1.58 nm in DCWF-6. When the pressure was too high, the Lc value of DCWF-8 increased to 1.21 nm. These results indicate that an appropriate hot-press densification treatment effectively regulates the interlayer spacing of the carbon layers and stacking thickness of the orderly arranged carbon layers along the c-axis in hard carbon materials. To better understand the effect of hot-press densification, the electrical conductivity of CWF and DCWF-P was measured using the four-probe method. Compared to CWF (9.8 S m⁻¹), the conductivity of DCWF-4, DCWF-6, and DCWF-8—samples obtained through hot-press densification—was increased to 17.4, 20.8, and 13.3 S m⁻¹, respectively. The structural characteristics of the CWF, DCWF-P and DCWF-x samples were analyzed using Raman spectroscopy. As Fig. 3b and S4b† show, all samples exhibited two characteristic peaks at 1344.4 and 1593.0 cm⁻¹, corresponding to the D-band and G-band of carbon materials, respectively. The D-band represents a disordered structure or defects in the carbon material, whereas the G-band indicates the in-plane vibration of sp² carbon atoms. The area ratio of the D-band to the G-band (I_D/I_G) can be used to characterize the degree of defects in the carbon materials.^31,32 By calculating the I_D/I_G values, it was found that the I_D/I_G values for CWF, DCWF-4, DCWF-6, DCWF-8, DCWF-20%, and DCWF-80% were 1.21, 1.33, 1.38, 1.28, 1.34, and 1.41, respectively. This suggests that moderate hot-press densification and increased initial moisture content in the WF contribute to a higher defect level in the carbon materials. Previous studies indicated that a higher defect level in carbon materials is closely associated with a richer closed-pore structure.³³

Fig. 3 Structure and composition characteristics of DCWF-P. XRD patterns (a), Raman spectra (b), and XPS spectra (c) of CWF and DCWF-P, respectively. C 1s XPS high-resolution spectra (d) of CWF and DCWF-P, respectively. N₂ adsorption/desorption isothermal curves (e) and pore distribution curves (f) of CWF and DCWF-P. SAXS curves (g) of CWF and DCWF-P. Closed pore diameter and fractal dimension fitted by SAXS (h).

XPS was used to analyze the elemental compositions and bonding states of the prepared samples. Fig. 3c and S5a† show the XPS survey spectra of CWF, DCWF-P and DCWF-x. All the samples exhibited only C 1s and O 1s peaks. Fig. 3d and S5b, c† present the high-resolution XPS spectra of C 1s for CWF, DCWF-P, DCWF-20%, and DCWF-80%. The high-resolution C 1s XPS profiles were deconvoluted into four different types of carbon configurations: sp² carbon (284.5 eV), sp³ carbon (285.5 eV), C–O (286.6 eV), and C=O (289.0 eV). Typically, the ratio of the fitted peak areas of sp³ carbon to sp² carbon reflects the degree of disorder in carbon materials.^34,35 Table S1† shows the fitting results for the high-resolution C 1s XPS spectra. Compared to the CWF (19.35%), the sp³ carbon content in DCWF-4, DCWF-6, DCWF-8, DCWF-20%, and DCWF-80% was reduced to 17.07%, 13.67%, 16.18%, 15.45%, and 12.04%, respectively. The decrease in sp³ carbon content was mainly due to the recrystallization of amorphous cellulose during the hot-press densification process. The higher crystallinity of cellulose allowed the formation of more cross-linked, disordered graphite-like layers during the high-temperature carbonization process. Specifically, the ratios of the fitting peak areas of sp² carbon to sp³ carbon for CWF, DCWF-4, DCWF-6, DCWF-8, DCWF-20%, and DCWF-80% were 2.67, 3.98, 5.04, 3.90, 3.59, and 5.33, respectively, indicating that the hot-press densification treatment is beneficial for increasing the degree of graphitization in CWF.

To analyze the pore structure of the materials, N₂ adsorption–desorption tests were conducted on the prepared samples. Fig. 3e shows the N₂ adsorption–desorption curves of the CWF and DCWF-P. According to the IUPAC classification, all samples exhibited adsorption hysteresis loops at medium relative pressures accompanied by capillary condensation phenomena, indicating that the isotherms belonged to type IV and that the hysteresis loops were of the H4 type. This suggests that the samples contain both mesopores and micropores.^36,37 Brunauer–Emmett–Teller equation calculations revealed that the specific surface areas of CWF, DCWF-4, DCWF-6, and DCWF-8 were 65.3, 55.6, 50.3, and 26.6 m² g⁻¹, respectively. Fig. 3f displays the pore size distribution curves for the CWF and DCWF-P. All samples show a rich and heterogeneous distribution of mesopores. Notably, when the processing pressure reached 8 MPa, the microporous structure of DCWF-8 decreased significantly. Studies have indicated that microporous structures enhance the specific surface area of a material, exposing more active sites, whereas mesoporous structures facilitate electrolyte penetration and ion/electron transfer.³⁸ Therefore, hot-press densification effectively regulates the specific surface area and pore structure of hard-carbon materials. However, due to the relatively large size of N₂ molecules, they have difficulty entering closed pores, which presents a limitation for the N₂ adsorption–desorption test in studying the closed pore structure of the samples. The pore size distribution of CWF and DCWF-6 was further analyzed using CO₂ adsorption–desorption test (Fig. S6a†). Compared to CWF, DCWF-6 exhibits a hierarchical ultramicroporous structure. The pore size distribution of DCWF-6 ranges from 0.54 to 0.82 nm, which is significantly larger than that of CWF. The hierarchical ultramicroporous structures enhance Na⁺ diffusion kinetics.

Small-angle X-ray scattering (SAXS) is an effective method for characterizing closed-pore structures. Closed pores are formed by highly disordered carbon layers, pseudo-graphitized structures, or graphitized-like structures, and their internal characteristics can be analyzed using the peak sizes and positions obtained from SAXS. The scattering vector Q in SAXS (Q = 4π sin θ/λ, where λ = 1.541 Å is the X-ray wavelength and θ is half the scattering angle) ranges from 1 to 2 nm⁻¹.^39,40Fig. 3g shows the SAXS spectra of CWF and DCWF-P. All samples exhibited a prominent peak at approximately 1–2 nm⁻¹, with the peak intensity of DCWF-P being higher than that of CWF. This indicates that the hot-press densification treatment favors an increase in both the diameter and content of the closed pores. When the processing pressure was excessively high, a reduction in the peak intensity of DCWF-8 was observed, suggesting that excessive pressure damaged the structure of WF, reduced the graphitic layers formed from carbonized crystalline cellulose, and decreased the closed pore structures formed through shrinkage encapsulation. The fitting results for the closed-pore diameter and fractal dimension obtained from the theoretical models, are shown in Fig. 3h. The average closed-pore diameter of the prepared samples was approximately 2 nm. With an increase in hot-press densification pressure, the closed pore diameter increases from 2.12 nm in CWF to 2.27 nm in DCWF-6. The surface roughness of the closed pore structures was evaluated in the high q range of the SAXS pattern using Porod fitting (Fig. S7†). The fitting results indicated that the surfaces of the closed pore structures becomes smoother with increasing hot-press densification pressure.¹⁵ To quantitatively analyze the characteristics of the closed-pore structure, the skeletal density of CWF and DCWF-6 samples was measured using a true density testing method based on Archimedes’ principle combined with the gas expansion displacement law (Fig. S6b†). This method employs helium as the measurement medium, which can penetrate all open pores except for the closed pores, thereby enabling precise determination of the skeletal density of the material. The results show that the skeletal density of directly carbonized wood fibers (CWF) reaches as high as 2.02 g cm⁻³, while the closed-pore volume is only 0.05 cm³ g⁻¹. After hot-press densification treatment, the skeletal density of DCWF-6 significantly decreases from 2.02 g cm⁻³ to 1.73 g cm⁻³, whereas the closed-pore volume increases from 0.05 cm³ g⁻¹ to 0.14 cm³ g⁻¹. This inverse trend between density and porosity suggests that hot-press densification effectively reconstructs the pore structure of hard carbon materials. In particular, the formation of closed nanopores in the DCWF-6 sample provides more active sites for sodium-ion storage in the low-voltage plateau region.

To investigate the effects of hot-press densification treatment on the pore structure formation mechanism of wood-fiber-derived hard-carbon materials, we analyzed WF and DWF pyrolyzed at different temperatures using XRD and electron paramagnetic resonance (EPR). Fig. 4a presents the XRD patterns of WF-T and DWF-T (where T represents the pyrolysis temperature) at various pyrolysis temperatures. From room temperature to 250 °C, the structure of the crystalline region of the cellulose remained essentially unchanged. In addition, when the pyrolysis temperature exceeds 300 °C, amorphous carbon peaks appear at 2θ = 18.9° for WF-300 and DWF-300, with the intensity of the amorphous carbon peak being stronger in DWF-300 than in WF-300. This indicates that DWF-300 generated more carbon microcrystals than WF-300 at the same temperature. When the pyrolysis temperature reaches 600 °C, a characteristic carbon (1 0 0) peak appears at 2θ = 44.1° in DWF-600, suggesting that DWF-600 undergoes stacking and carbonization earlier than WF-600. Notably, as the pyrolysis temperature increased, the (0 0 2) characteristic peaks of the WF-T and DWF-T samples shifted positively, indicating a reduction in carbon layer spacing. When the pyrolysis temperature reaches 700 °C, the positive shift of the (0 0 2) characteristic peaks in the WF-700 and DWF-700 samples slows down.⁴¹

Fig. 4 Pore structure formation mechanism of DCWF-6. XRD patterns (a) of WF and DWF at different calcination temperatures. EPR spectra of WF and DWF carbonized at 300 °C (b) and 1300 °C (c), respectively.

EPR was used to characterize radical concentrations and types in different samples. Fig. 4b shows the EPR spectra of WF-300 and DWF-300. The figure shows that the spectral shapes of WF-300 and DWF-300 are essentially the same, but their intensities and positions exhibit significant differences, indicating that the hot-press densification process affected the radicals in the WFs. In addition, during pyrolysis at lower temperatures, a large number of radicals did not participate in the rearrangement of the carbon framework in the hard carbon. When the pyrolysis temperature increased to 1300 °C, the radicals in WF-1300 and DWF-1300 became more delocalized (Fig. 4c), and the stacking between graphite microcrystals led to the formation of a closed-pore structure.^42,43

To investigate the effects of the hot-press densification treatment on the WF pyrolysis mechanism, large-scale molecular dynamics simulations were conducted. Both pristine bilayer and defective graphene structures were constructed. For defective graphene, typical double-vacancy defects and Stone–Wales defects were introduced, with the defect locations staggered between the two layers. The initial interlayer distances for pristine bilayer and defective graphene were set to 0.39 and 0.36 nm, respectively, with a total of 768 and 764 carbon (C) atoms. In addition, 100 isolated C atoms were randomly distributed 0.3 nm above and below the graphene layer (200 discrete C atoms per model). The initial model structure is illustrated in Fig. S8.† At relatively low temperatures, numerous isolated carbon atoms rapidly bind epitaxially to graphene layers. In comparison, the defective graphene structure exhibited a denser arrangement, potentially due to the reduced interlayer spacing and lower surface area, which are characteristics of hard carbon prepared through hot-press densification. As the temperature increases, these isolated carbon atoms gradually disperse to form a sparse system with graphene layers oriented in random directions (Fig. 5a and b). Dynamic visualizations of these two processes are provided in the ESI (Movies S1 and S2†). As the temperature increases, the bonding between the isolated carbon atoms and graphene gradually intensifies (Fig. S9†). Fig. 5c shows the sp²/sp³ ratios of the two models during carbonization. The sp²/sp³ ratio in the defective graphene changed more slowly and was higher than that in the pristine graphene, indicating that hot-press densification promoted the formation of more graphite-like layers. In addition, analysis of the interlayer spacing changes in both models during carbonization revealed that the interlayer spacing decreased gradually at lower temperatures but increased as the temperature increased (Fig. 5d). Notably, the defective graphene structure exhibited a larger interlayer spacing, likely because the defect sites acted as active centers, hindering the orderly rearrangement of certain carbon atoms, thereby expanding the interlayer spacing within the carbon matrix. Throughout the carbonization process, the defect structures vary with temperature, supporting their role as critical factors in the formation of closed-pore structures.

Fig. 5 Molecular dynamics simulations. Pristine graphene (a) and defective graphene (b) structures. Change of the structural composition for materials under different conditions. The sp²/sp³ ratio for the two models during carbonization (c). The interlayer spacing changes of both models during carbonization (d).

Thus, an increase in the crystalline cellulose content in WF through hot-press densification pretreatment results in a higher crystalline cellulose content that generates more cross-linked disordered graphite-like layers during high-temperature carbonization. Concurrently, lignin cross-linking and partial hemicellulose decomposition during hot-press densification yield a carbon matrix with more defective structures. Under high-temperature conditions, these defect structures serve as active centers that limit the ordered rearrangement of carbon atoms, facilitating the formation of numerous closed-pore structures in the hard carbon.

3.2 Electrochemical performances of half-cells

CWF and DCWF-P were assembled into coin-type half-cells in a glove box to test their sodium storage performance as anode materials. Fig. 6a shows the CV curves of CWF and DCWF-P. It was observed that the CV curves of the samples exhibit a distinct pair of redox peaks in the voltage range of 0.01–0.20 V, corresponding to the plateau regions of the charge–discharge curves. In the voltage range of 0.20–1.25 V, there is a broad current region corresponding to the sloped parts of the charge–discharge curves (Fig. 6b). Notably, the CV curve of DCWF-6 also shows the smallest irreversible area (Fig. S10†), indicating that the DCWF-6 electrode undergoes less dielectric decomposition, which is beneficial for improving the coulombic efficiency of the battery.^44,45 In addition, the first-cycle CV curve of the DCWF-6 electrode aligns well with subsequent scans, demonstrating excellent cycling stability and a thin SEI. Fig. 6b shows the initial galvanostatic charge–discharge (GCD) curves of the CWF and DCWF-P electrodes. The shapes of the GCD curves of the prepared electrodes were consistent, all displaying a high-potential sloped capacity and a low-potential plateau capacity with long plateau characteristics, which are typical of hard-carbon sodium storage.⁴⁶ The initial discharge specific capacities of CWF, DCWF-4, DCWF-6, and DCWF-8 were 378.5, 419.7, 427.1, and 386.4 mA h g⁻¹, respectively. The ICEs of the CWF, DCWF-4, DCWF-6, and DCWF-8 were 65.1%, 72.4%, 86.4%, and 72.0%, respectively. Fig. 6c summarizes the plateau and sloped specific capacities of the CWF and DCWF-P electrodes during the second discharge cycle. It can be seen that DCWF-6 has the highest plateau and sloped specific capacities. Combined with the structural characterization, the high plateau specific capacity of DCWF-6 is mainly attributed to its structure, which is rich in closed pores. These closed pores are believed to be sodium storage sites that provide plateau capacity. Fig. 6d shows the rate performance of the CWF and DCWF-P electrodes. The reversible specific capacities of the DCWF-6 electrode at current densities of 0.1, 0.2, 0.5, 1.0, 2.0, and 4.0 A g⁻¹ are 427.1, 357.7, 340.9, 313.5, 277.5, and 197.7 mA h g⁻¹, respectively, outperforming CWF, DCWF-4, and DCWF-8. When the current density returns to 0.1 A g⁻¹, the reversible specific capacity of the DCWF-6 electrode reaches 361.4 mA h g⁻¹, indicating good reversibility. To further demonstrate the structural stability of DCWF-6 during sodium storage, cycling performance tests were conducted at a current density of 0.2 A g⁻¹. As Fig. 6e shows, the discharge specific capacity of the DCWF-6 electrode was higher than those of the CWF, DCWF-4, and DCWF-8 electrodes, and after 100 charge–discharge cycles, the discharge specific capacity remained at 362.0 mA h g⁻¹ with a coulombic efficiency close to 100%, indicating excellent structural stability during sodium storage. In addition, after 300 charge–discharge cycles at a high current density of 0.5 A g⁻¹, the capacity retention of the DCWF-6 electrode remained at 84.8% (Fig. 6f). To investigate the effect of initial water content on the electrochemical performance of DCWF, the rate capability and cycle stability of DCWF-x were tested. Fig. S11a† shows the GCD curves of DCWF-x. The initial discharge specific capacities of DCWF-20% and DCWF-80% were 398.1 and 404.7 mA h g⁻¹, respectively. Notably, during the rate performance testing of DCWF-x (Fig. S11b†), it was found that when the current density was increased to 4.0 A g⁻¹, the discharge specific capacity of the DCWF-20% electrode was slightly lower than that of the CWF electrode. This may be due to the slower sodium storage kinetics in the closed pores, where an excessive amount of closed pore structure could negatively impact the rate performance. In addition, all DCWF-x electrodes exhibited good cycle stability (Fig. S11c†). Compared with other previously reported biomass-derived hard-carbon materials (Fig. 6g), DCWF-6 shows significant advantages in terms of rate performance and reversible specific capacity.^{1,8,15,16,37,47–51}

Fig. 6 Sodium anodic performance of CWF and DCWF-P. CV curves (a) of CWF and DCWF-P at the scan rate of 0.5 mV s⁻¹. The first discharge–charge curves (b) of CWF and DCWF-P at the current density of 0.1 A g⁻¹. Capacity distributions (c) of CWF and DCWF-P at different potential region. Rate performance (d) of CWF and DCWF-P. Cycling performance (e) of CWF and DCWF-P at the current density of 0.2 A g⁻¹. Cycling performance (f) of DCWF-6 at the current density of 0.5 A g⁻¹. Comparison of capacities at different current densities with other previously reported biomass-derived hard carbon anodes (g).

The sodium storage mechanism of the DCWF-6 electrode was analyzed by in situ Raman spectroscopy, as shown in Fig. 7a. The in situ Raman spectrum of the DCWF-6 electrode exhibits strong D- and G-bands at 1350 and 1600 cm⁻¹, respectively. During the initial discharge process in the voltage range of 2–0.01 V, the peak intensity of the D-band gradually decreased, whereas the peak position of the G-band redshifted and the peak shape broadened. Previous studies have shown that the gradual intercalation of sodium ions into the carbon layers and the uniform diffusion of sodium ions, leading to electronic doping, cause changes in the G peak position. After discharging to 0.1 V, the G-band splits into two peaks, which gradually intensify as the voltage decreases. These results indicate the formation of a multiphase structure due to sodium intercalation into the carbon layers. The characteristic peak at 1610 cm⁻¹ is associated with the Na-GIC boundary layer, whereas the peak at 1475 cm⁻¹ represents electronic doping caused by sodium ion intercalation into the carbon layers. Previous studies have indicated that when sodium intercalates into the graphene gaps to form NaC₂₄, it reaches saturation. If more sodium continues to intercalate into the graphite gaps, the G-band continues to shift to lower frequencies.^2,52 Therefore, when DCWF-6 forms a complex carbon structure, sodium ions continue to intercalate. In addition, the redshift of the G peak position due to the formation of sodium clusters recovered well during charging, confirming the stability of the DCWF-6 structure.

Fig. 7 Sodium-ion storage mechanism and structure–performance relationship. In situ Raman spectra of DCWF-6 (a). b-Values of CWF and DWCF-P at 0.5 V (b) and at 0.1 V (c). The GITT potential profiles of CWF and DWCF-P for the sodiation process (d) and the desodiation process (e). EIS profiles (f) of CWF and DWCF-P. Ex situ EIS profiles of DCWF-6 in discharged (g) and charged (h) states during the initial cycle.

To gain a deeper understanding of the electrochemical kinetics of the prepared electrodes, CV and galvanostatic intermittent titration (GITT) tests were performed on the CWF and DCWF-P electrodes at different scan rates. Fig. S12a, S13a, S13d and S13g† show the CV curves of the CWF and DCWF-P electrodes at various scan rates. The CWF and DCWF-P electrodes exhibited similar curve shapes, indicating that their sodium storage mechanisms were fundamentally consistent. The relationship between the scan rate (v) and current density (i) satisfies i(v) = av^b (where a and b are constants), which can be used to qualitatively analyze the sodium storage mechanism of DCWF-6. When b = 0.5, it indicates a diffusion-controlled charge storage process, and when b = 1.0, it indicates a capacitively controlled charge storage process.⁵⁰ According to the linear fit (log i–log v), the b values of the CWF, DCWF-4, DCWF-6, and DCWF-8 electrodes at 0.5 V are 0.95, 0.96, 1.02, and 0.93, respectively (Fig. 7b). The b values of the prepared electrodes were close to 1.0, indicating that the sloped region contained abundant surface-defect adsorption active sites, and that the sodium-ion storage mechanism was controlled by the capacitive effect. In addition, the b values of the CWF, DCWF-4, DCWF-6, and DCWF-8 electrodes at 0.05 V are close to 0.5 (Fig. 7c), indicating that the plateau region contains closed-pore structures and that the sodium ion storage mechanism is diffusion-controlled. These results confirmed that the sodium storage process of the CWF and DCWF-P samples followed an adsorption and pore-filling model. To quantify the diffusion/capacitance contribution ratio, the formula i(v) = k₁v + k₂v^1/2 was used: Fig. S12 and S13† compare the pseudocapacitive contributions of CWF, DCWF-4, DCWF-6, and DCWF-8 electrodes. The results showed that as the scan rate increased, the capacitance contribution ratio gradually increased.⁵³ At a scan rate of 1.0 mV s⁻¹, the capacitance contributions of the CWF, DCWF-4, DCWF-6, and DCWF-8 electrodes were 57.21%, 61.41%, 63.19%, and 59.78%, respectively. The DCWF-6 electrode exhibited the highest capacitance contribution. This capacitive contribution, resulting from the closed-pore structure regulated by hot-press densification, enables rapid sodium ion storage in hard carbon materials under high-current discharge conditions.

The GITT curves of the CWF and DCWF-P electrodes are presented in Fig. S14.† All the electrodes exhibited similar trends in their GITT curves (Fig. 7d and e). As Na⁺ intercalated into the hard-carbon phase, the diffusion coefficient of Na⁺ slowly decreased, then significantly increased, and finally began to increase toward the end of the discharge process. Lower diffusion coefficients were mainly distributed in the low-voltage region, which was attributed to Na⁺ preferentially adsorbing on the edges of the graphite layers and surface defects in the sloping region (above 0.1 V), demonstrating excellent kinetic characteristics. In the plateau region (below 0.1 V), Na⁺ filling the nanopores required more energy to overcome the electrostatic repulsion caused by the charge distribution, resulting in slower diffusion of Na⁺ in the bulk phase. These results are consistent with those of the in situ Raman and CV curves obtained at different scan rates. In addition, electrochemical impedance spectroscopy (EIS) tests were conducted to analyze the Na⁺ transport properties in the battery. Fig. 7f shows Nyquist plots of the CWF and DCWF-P electrodes. All electrodes primarily consisted of a sloped line in the low-frequency region and a semicircle in the high-frequency region. The sloped line at low frequencies indicates the diffusion resistance (W₀) of the solid material, whereas the semicircle in the high-frequency region represents the charge-transfer resistance (R_ct). The figure shows that the R_ct value of the DCWF-6 electrode (2.79 Ω) is significantly lower than those of CWF (3.90 Ω), DCWF-4 (3.41 Ω), and DCWF-8 (3.82 Ω), indicating that DCWF-6 has a lower charge transfer resistance. In addition, the slope in the low-frequency region was the highest for DCWF-6, indicating that it had the lowest diffusion resistance. Considering the changes in charge transfer kinetics at different voltages, in situ EIS analysis was performed on DCWF-6. The fitting results for R_ct during the sodiation/desodiation processes are shown in Fig. 7g and h. The figures clearly show that R_ct increased slightly during the discharge process, mainly because of Na⁺ insertion into the graphite layers.⁵⁴ These results demonstrate that the DCWF-6 electrode has the lowest charge transfer resistance at the electrode–electrolyte interface during the sodiation/desodiation process, further proving that an appropriate thermal compaction treatment can promote the formation of DCWF-6.

3.3 Electrochemical performances of full-cells

To investigate the practical application performance of the DCWF-6 electrode, a full cell was assembled using pre-sodiated DCWF-6 as the anode and Na₃V₂(PO₄)₃ (NVP) as the cathode, with an anode-to-cathode mass ratio of 1 : 2.5 (Fig. S15a†). The DCWF-6//NVP battery exhibited a high reversible specific capacity of 284.9 mA h g⁻¹ at 30 mA g⁻¹ (based on the anode mass), with an average operating voltage of 3.21 V (Fig. S15b†). Furthermore, as the current density increased to 30, 50, 100, 200, and 400 mA g⁻¹, the specific capacities of the DCWF-6//NVP battery were 284.9, 236.6, 217.7, and 186.0 mA h g⁻¹, respectively (Fig. S15c†), indicating excellent rate performance and fast charging capability. When the current density returned to 30 mA g⁻¹, the capacity exhibited excellent reversibility, reaching 218.9 mA h g⁻¹. Fig. S15d† reveals the long-term cycling performance of the DCWF-6//NVP battery at 30 mA g⁻¹. After 100 charge–discharge cycles, the DCWF-6//NVP battery maintained a high specific capacity of 167.5 mA h g⁻¹. Based on the total mass of the active materials in the anode and cathode, the energy density of the DCWF-6//NVP battery was 224.6 W h kg⁻¹, demonstrating the significant application potential of DCWF-6 in practical SIBs.

4. Conclusions

The composition and microstructure of WF were effectively regulated through a simple thermal compression densification treatment. During hot-press densification, amorphous cellulose undergoes recrystallization, which increases the crystalline cellulose content and enhances the stability of the internal fiber structure. Simultaneously, partial degradation of hemicellulose results in the generation of highly active free radicals. These free radicals play a crucial role in the formation of the CWF microstructures. The high crystalline cellulose content in the WF promotes the formation of cross-linked disordered graphite-like layers during carbonization. These graphite-like layers are then curled and folded to form closed pores, further strengthening the structure of the material. Hot-press densification treatment facilitates the formation of closed-pore structures, increases the interlayer spacing of the carbon layers, and reduces the specific surface area of the CWF. In this study, the carbonization mechanism of the WF was demonstrated by molecular dynamics simulations. Results showed that the densified CWF exhibited improved ICE. Compared to untreated CWF, the densified WFs (DCWF-6) demonstrated superior sodium storage performance, achieving a high reversible specific capacity of 427.1 mA h g⁻¹ at a current density of 0.1 A g⁻¹, with a plateau capacity of 242.4 mA h g⁻¹ and an ICE of 86%. The exceptional electrochemical performance of DCWF-6 could be attributed to its unique microstructure, which includes a lower specific surface area, larger interlayer spacing, and an abundant closed-pore structure. This simple and effective thermal compression densification method has great potential for improving the performance of SIBs and can be applied to other biomass precursors for enhanced energy storage applications.

Author contributions

Yangyang Chen: conducting experiment, data curation, writing – original draft preparation. Yu Liao: formal analysis, investigation, validation. Yiding Ding: formal analysis, investigation. Ying Wu: data curation. Lei Li: formal analysis. Sha Luo: methodology. Yan Qing: writing – reviewing and editing. Zhihan Li: supervision. Zhen Zhang: investigation. Yiqiang Wu: writing – reviewing and editing. All authors read and approved the final manuscript.

Data availability

The data supporting this article have been included as part of the ESI.† The data that support the findings of this study are available from the corresponding author, Y. Qing, upon reasonable request.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was financially supported by the National Key Research and Development Program of China (2023YFD2200503), the Science and Technology Innovation Program of Hunan Province (2022RC3054).

References

1. Z. Tang, R. Zhang, H. Wang, S. Zhou, Z. Pan, Y. Huang, D. Sun, Y. Tang, X. Ji, K. Amine and M. Shao, Nat. Commun., 2023, 14, 6024.
2. J. Zhao, X.-X. He, W.-H. Lai, Z. Yang, X.-H. Liu, L. Li, Y. Qiao, Y. Xiao, L. Li, X. Wu and S.-L. Chou, Adv. Energy Mater., 2023, 13, 2300444.
3. Y. Zhao, Y. Kang, J. Wozny, J. Lu, H. Du, C. Li, T. Li, F. Kang, N. Tavajohi and B. Li, Nat. Rev. Mater., 2023, 8, 623.
4. Y. Chu, J. Zhang, Y. Zhang, Q. Li, Y. Jia, X. Dong, J. Xiao, Y. Tao and Q.-H. Yang, Adv. Mater., 2023, 35, 2212186.
5. F. Wang, Z. Jiang, Y. Zhang, Y. Zhang, J. Li, H. Wang, Y. Jiang, G. Xing, H. Liu and Y. Tang, eScience, 2024, 4, 100181.
6. A. Rudola, R. Sayers, C. J. Wright and J. Barker, Nat. Energy, 2023, 8, 215.
7. B. Chen, L. Zhong, M. Lu, W. Jian, S. Sun, Q. Meng, T. Wang, W. Zhang and X. Qiu, Green Chem., 2024, 26, 7919–7930.
8. K. Wang, F. Sun, H. Wang, D. Wu, Y. Chao, J. Gao and G. Zhao, Adv. Funct. Mater., 2022, 32, 2203725.
9. J. Liu, Y. You, L. Huang, Q. Zheng, Z. Sun, K. Fang, L. Sha, M. Liu, X. Zhan, J. Zhao, Y.-C. Han, Q. Zhang, Y. Chen, S. Wu and L. Zhang, Adv. Mater., 2024, 36, 2407369.
10. X. Feng, Y. Li, Y. Li, M. Liu, L. Zheng, Y. Gong, R. Zhang, F. Wu, C. Wu and Y. Bai, Energy Environ. Sci., 2024, 17, 1387.
11. B. Zhong, C. Liu, D. Xiong, J. Cai, J. Li, D. Li, Z. Cao, B. Song, W. Deng, H. Peng, H. Hou, G. Zou and X. Ji, ACS Nano, 2024, 18, 16468.
12. F. Luo, C. Yi, D. Yang, D. Fan, W. Liu, X. Qiu and W. Zhang, Green Chem., 2025, 27, 1703–1713.
13. Y. Chen, Y. Liao, Y. Wu, L. Li, Z. Zhang, S. Luo, Y. Wu and Y. Qing, J. Energy Chem., 2024, 95, 701.
14. Y. Chen, Y. Liao, Y. Qing, Y. Ding, Y. Wu, L. Li, S. Luo and Y. Wu, J. Energy Storage, 2024, 99, 113186.
15. Y. Wang, Z. Yi, L. Xie, Y. Mao, W. Ji, Z. Liu, X. Wei, F. Su and C.-M. Chen, Adv. Mater., 2024, 36, 2401249.
16. W. Zhang, Z. Huang, H. N. Alshareef and X. Qiu, Carbon Res., 2024, 3, 28.
17. J. Gao, X. Wang, Q. Xu, B. Kuai, Z. Wang, L. Cai, S. Ge, Y. L. Zhang and G. Li, Ind. Crops Prod., 2023, 193, 116291.
18. J. Kang, F. Yang, C. Sheng, H. Xu, J. Wang, Y. Qing, Y. Wu and X. Lu, Small, 2022, 18, 2200950.
19. A. B. Vanzetto, L. V. R. Beltrami and A. J. Zattera, Cellulose, 2021, 28, 6967.
20. Y. Mao, Z. Yi, L. Xie, L. Dai, F. Su, Y. Wang, W. Ji, X. Wei, G. Hui, Y. Chang, W. Xie, G. Sun, D. Jiang and C.-M. Chen, Energy Storage Mater., 2024, 73, 103845.
21. H. Alqrinawi, B. Ahmed, Q. Wu, H. Lin, S. Kameshwar and M. Shayan, Ind. Crops Prod., 2024, 213, 118430.
22. G. Xia, S. Yu, Z. Yang, J. Lian, Y. Xie and Q. Feng, Int. J. Biol. Macromol., 2024, 281, 136187.
23. H. Luo, R. Si, J. Liu, P. Li, Y. Tao, X. Zhao and H. Chen, Cellulose, 2022, 29, 7377.
24. D. Cheng, H. Ding, B. Xu and T. Li, Ind. Crops Prod., 2024, 221, 119376.
25. M. Jakob, J. Janesch, U. Müller and W. Gindl-Altmutter, ACS Sustainable Chem. Eng., 2024, 12, 16800.
26. Y. Ran, D. Lu, J. Jiang, Y. Huang, W. Wang and J. Cao, Chem. Eng. J., 2023, 471, 144476.
27. M. Chai, L. Xie, X. Yu, X. Zhang, Y. Yang, M. M. Rahman, P. H. Blanco, R. Liu, A. V. Bridgwater and J. Cai, Renewable Sustainable Energy Rev., 2021, 143, 110962.
28. X. Li, S. Zhang, J. Tang, J. Yang, K. Wen, J. Wang, P. Wang, X. Zhou and Y. Zhang, J. Mater. Chem. A, 2024, 12, 21176.
29. Y. Chen, Y. Wu, L. Li, Y. Liao, S. Luo, Y. Wu and Y. Qing, Chem. Eng. J., 2023, 461, 141988.
30. Y. Zeng, J. Yang, H. Yang, Y. Yang and J. Zhao, ACS Energy Lett., 2024, 9, 1184.
31. L. Xie, C. Tang, Z. Bi, M. Song, Y. Fan, C. Yan, X. Li, F. Su, Q. Zhang and C. Chen, Adv. Energy Mater., 2021, 11, 2101650.
32. X. Li, J. Sun, W. Zhao, Y. Lai, X. Yu and Y. Liu, Adv. Funct. Mater., 2022, 32, 2106980.
33. Y. Li, Y. Lu, Q. Meng, A. C. S. Jensen, Q. Zhang, Q. Zhang, Y. Tong, Y. Qi, L. Gu, M.-M. Titirici and Y.-S. Hu, Adv. Energy Mater., 2019, 9, 1902852.
34. H. Zhang, W. Zhang and F. Huang, Chem. Eng. J., 2022, 434, 134503.
35. D. Cheng, Z. Li, M. Zhang, Z. Duan, J. Wang and C. Wang, ACS Nano, 2023, 17, 19063.
36. P. Zheng, W. Zhou, Y. Mo, B. Zheng, M. Han, Q. Zhong, W. Yang, P. Gao, L. Yang and J. Liu, J. Energy Chem., 2025, 100, 730.
37. J. Yang, X. Wang, W. Dai, X. Lian, X. Cui, W. Zhang, K. Zhang, M. Lin, R. Zou, K. P. Loh, Q.-H. Yang and W. Chen, Nano-Micro Lett., 2021, 13, 98.
38. W. Zhao, H. Dong, Z. Xing, L. Zhou, S. Chen, H.-K. Liu, S.-X. Dou, Z. Zhao, H. Xia, S. Chou and M. Chen, Adv. Energy Mater., 2024, 14, 2402720.
39. A. Kamiyama, K. Kubota, D. Igarashi, Y. Youn, Y. Tateyama, H. Ando, K. Gotoh and S. Komaba, Angew. Chem., Int. Ed., 2021, 60, 5114.
40. H. Au, H. Alptekin, A. C. S. Jensen, E. Olsson, C. A. O'Keefe, T. Smith, M. Crespo-Ribadeneyra, T. F. Headen, C. P. Grey, Q. Cai, A. J. Drew and M.-M. Titirici, Energy Environ. Sci., 2020, 13, 3469.
41. K. Kupwade-Patil, S. D. Palkovic, A. Bumajdad, C. Soriano and O. Büyüköztürk, Constr. Build. Mater., 2018, 158, 574.
42. F. Khelifa, S. Ershov, Y. Habibi, R. Snyders and P. Dubois, Chem. Rev., 2016, 116, 3975.
43. G. Liu, Z. Wang, H. Yuan, C. Yan, R. Hao, F. Zhang, W. Luo, H. Wang, Y. Cao, S. Gu, C. Zeng, Y. Li, Z. Wang, N. Qin, G. Luo and Z. Lu, Adv. Sci., 2023, 10, 2305414.
44. J. Jiao, C. Yi, X. Qiu, D. Yang, F. Fu and W. Liu, Green Chem., 2024, 26, 6643–6655.
45. Z. Zheng, S. Hu, W. Yin, J. Peng, R. Wang, J. Jin, B. He, Y. Gong, H. Wang and H. J. Fan, Adv. Energy Mater., 2024, 14, 2303064.
46. J. M. Stratford, A. K. Kleppe, D. S. Keeble, P. A. Chater, S. S. Meysami, C. J. Wright, J. Barker, M.-M. Titirici, P. K. Allan and C. P. Grey, J. Am. Chem. Soc., 2021, 143, 14274.
47. X. Zhang, Y. Cao, G. Li, G. Liu, X. Dong, Y. Wang, X. Jiang, X. Zhang and Y. Xia, Small, 2024, 20, 2311197.
48. T. Xu, X. Qiu, X. Zhang and Y. Xia, Chem. Eng. J., 2023, 452, 139514.
49. K. Wang, M. Li, Z. Zhu, W. Ai, H. Wu, B. Wang, P. He, D. Xie, J. Wu and W. Huang, Nano Energy, 2024, 124, 109459.
50. Z. Guo, Z. Xu, F. Xie, J. Jiang, K. Zheng, S. Alabidun, M. Crespo-Ribadeneyra, Y.-S. Hu, H. Au and M.-M. Titirici, Adv. Mater., 2023, 35, 2304091.
51. K. Yu, X. Wang, H. Yang, Y. Bai and C. Wu, J. Energy Chem., 2021, 55, 499.
52. C. Cai, Y. Chen, P. Hu, T. Zhu, X. Li, Q. Yu, L. Zhou, X. Yang and L. Mai, Small, 2022, 18, 2105303.
53. S. Chong, S. Qiao, Z. Wang, B. Chen, B. Yuan, W. Huang and G. Cao, Nano Energy, 2024, 130, 110097.
54. H. Zhang, F. Gao, D. Zhang, C. Gao, G. Huang, Z. Zhang, Y. Liu, M. Terrones, J. Wei and Y. Wang, Carbon, 2024, 230, 119602.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d5gc00409h

---