Introduction

Rapid development of new energy transportation and the need for power sources operating in extreme environments, such as aerospace and submarine equipment, spotlights the safety of electrical energy storage systems1,2,3. Supercapacitors, as burgeoning storage devices, are no exception4,5,6. Their rapid charge-discharge rates render them suitable for extreme environments, such as engine ignition systems that require high instantaneous power, demanding high working stability and damage resistance. Conventionally, energy storage devices are safeguarded with protective shields, which compromise weight and space efficiency while offering protection, potentially increasing energy self-consumption7,8. Moreover, inner components like electrodes and electrolytes still remain susceptible to damage. It is imperative to bolster the resilience of supercapacitors’ core components without forfeiting lightness and space efficiency. Additionally, ensuring the scalability and adaptability of these reinforced supercapacitors is crucial for mass production and broadening their application range9,10.

In supercapacitor devices, the operational conditions are largely dictated by the core components, including electrodes and electrolytes11,12. Firstly, ensuring damage resistance in electrodes using load-bearable materials is paramount, enhancing both longevity and reliability, particularly in harsh environments such as pressure loading13,14,15. Polymer-derived ceramic (PDC) silicon oxycarbide (SiOC) stands out due to its robust physical and chemical properties16,17,18, exhibiting impressive compressive stress of up to 70 MPa and energy absorption capacities reaching 100 kJ/cm3. Furthermore, the conductivity of SiOC can be easily tailored through methods such as high-temperature graphitization, surface coatings, and metal doping, positioning it as a potential candidate for current collectors19,20,21. When coupled with active materials, this robust composite presents an avenue for reinforced supercapacitor electrode fabrication22. Since electrodes are typically engineered into specific configurations and dimensions to fully utilize their electrochemical and mechanical performance in device assemblies, a further significant advantage of SiOC substrates is their compatibility with three-dimensional (3D) printing techniques for flexible configuration design. This adaptability facilitates the optimal use of confined spaces by electrode architecture adjustment, maximizing active material load and minimizing weight, thereby reducing energy self-consumption, while fine-tuning the mechanical properties23,24,25. Consequently, the employment of 3D-printed SiOC lattice-based electrodes is promising to enhance the reliability of the supercapacitor device by taking advantage of their optimized mechanical behavior and space utilization26,27.

Secondly, the choice of damage-invulnerable electrolyte for stable ion exchange, merits keen consideration for further improving supercapacitor safety28,29,30. Conventional liquid electrolytes often exert pressure on electrodes and risk leakage upon device compromise. Thus, hydrogel electrolytes emerge as a promising alternative for supercapacitor devices31,32,33. They not only imbibe electrolytes due to inherent hydrophilicity and porousness, rivaling the performance of liquid electrolytes, but their crosslinked microstructures also afford notable flexibility, enabling them to endure extensive deformation without function loss34. Additionally, molecular chain modifications can bestow hydrogels with self-healing properties, further mitigating damage vulnerability. Flexible supercapacitors employing hydrogel electrolytes, exhibit considerable potential for use under complex load-bearing conditions, distinguished by their high impact tolerance and superior electrochemical performance35,36,37. Meanwhile, rigid and impact-resistant supercapacitors are also essential in specific applications such as aerospace, underwater equipment, and new energy vehicles38. Consequently, enhancing hydrogel supercapacitors with 3D-printed SiOC lattice-based electrodes holds promise for broadening their application scope. The hydrogel can be fabricated starting from a liquid solution in order to penetrate the internal space of the 3D-printed SiOC electrodes, thereby ensuring comprehensive contact. Subsequent crosslinking can transform the solution into a hydrogel, establishing a firm and cohesive interface with the electrodes. Consequently, integrating hydrogel electrolytes with 3D-printed SiOC lattice-strengthened electrodes is promising to yield practical supercapacitors exhibiting enhanced mechanical stability and diminished damage susceptibility39,40.

The assembly strategy of the supercapacitor cell is also crucial to mechanical stability. Given that supercapacitors employ an electrode-electrolyte-electrode configuration, they are typically constructed using lamellar structures. This arrangement may introduce mechanical weak points determined by the least robust component within the structure and induce anisotropy in mechanical behavior. To develop a mechanically coherent and isotropic supercapacitor cell, we employed a continuous all-SiOC-based skeleton to enhance mechanical performance uniformly. Triply periodic minimal surface (TPMS) configurations were used to construct the SiOC skeleton, leveraging their structural uniformity and periodicity. Electrodes were fabricated via in situ growth of active materials on the TPMS-structured SiOC, paired with an insulating SiOC separator that shared the same TPMS design. The hydrogel electrolyte was infused into the hollow spaces to complete a hydrogel-infused TPMS lattice configuration as a ready-to-use supercapacitor cell. Since this design was reinforced by the continuous SiOC skeleton, weak phases and anisotropic mechanical behavior common in lamellar or sandwich structures can be avoided. This uniform continuity and periodicity between the electrode and separator allowed the skeleton to distribute stress evenly, mitigating stress concentration. Furthermore, the employment of scalable and customizable 3D printing techniques endowed this strategy with the potential for creating comprehensive, tailored, damage-resistant, and ready-to-use supercapacitors.

Consequently, we provide a strategy for damage-invulnerable and ready-to-use supercapacitor cell fabrication, realized by polyvinyl alcohol (PVA) hydrogel electrolytes infused 3D-printed SiOC lattice electrodes (Fig. 1a). Firstly, the SiOC substrates were 3D printed into TPMS structures, notably gyroid (G), diamond (D), primitive (P), and I-wrapped package (I-WP), chosen for their mechanical robustness and periodicity41. Utilizing the precision and high efficiency of vat photopolymerization (VP), rapid and accurate formations can be achieved using a digital light processing (DLP) printer42,43. TPMS SiOC was then coated by pyrolytic carbon (C/SiOC), given its conductive, anti-corrosive, and lightweight attributes, in order to enhance conductivity for serving as robust current collectors. Secondly, we selected commercially viable polyaniline (PANI) as the energy storage material due to its features of high conductivity, adhesion, and lightness, and integrated it in situ onto C/SiOC to yield reinforced electrodes (PANI/C/SiOC)44. PANI demonstrates the capability to store or release charges via transitions among various redox states. Its facile formation of nanostructures on substrate materials contributed to an increase in surface roughness and the formation of nanopores, which collectively enlarged the accessible surface areas of the 3D electrodes. Moreover, it exhibited high compatibility with hydrogel electrolytes, facilitating efficient ion exchange, while adhering securely to the current collectors without using binders that ensured a low-loss electron exchange and prevented detachment. Finally, PVA hydrogel solution with dissolved H2SO4, was injected into molds printed via fused deposition modeling (FDM). These molds contained a pair of PANI/C/SiOC electrodes sandwiched by a SiOC separator sharing the same TPMS structure. Subsequent crosslinking of the solution resulted in the formation of hydrogel electrolytes, integrating these components. Upon demolding, self-standing and independently functioning supercapacitor cells were produced. By varying TPMS configurations subsequently, specific capacitance and mechanical performance of the cells were fine-tuned. Their post-impact electrochemical properties were examined, followed by an assessment of their performance under dynamic pressure conditions. Additionally, the self-healing capabilities of the hydrogel were investigated, with a subsequent evaluation of its electrochemical properties after the healing process. Furthermore, we explored scalability, aiming to provide a universal fabrication strategy.

Fig. 1: Schematics of the hydrogel-infused lattice supercapacitor and mechanical performance.
figure 1

a Schematic illustration of the polyvinyl alcohol (PVA)-H2SO4 hydrogel-infused polyaniline (PANI)/C/SiOC lattice supercapacitor. b Photograph showing the 3D-printed preforms, C/SiOC current collectors, and PANI/C/SiOC electrodes. c XRD patterns of the SiOC, C/SiOC, and PANI/C/SiOC. d Photograph showing the PANI/C/SiOC cells. e Density of the PANI/C/SiOC electrodes and PANI/C/SiOC cells. Data were presented as mean ± standard deviation from n = 5 independent samples. f Compressive stress curves of the PANI/C/SiOC cells with different TPMS structures, insert showing the lighting of a light-emitting diode under pressure loading of 10 MPa. g Compressive stress, Young’s modulus, and energy absorption values of PANI/C/SiOC cells with different TPMS structures. Data were presented as mean ± standard deviation from n = 5 independent samples.

Full size image

Results

Configuration design of hydrogel-infused lattice supercapacitor

The fabrication of PANI/C/SiOC lattice electrodes commenced with the 3D printing of a photocurable-modified SiOC precursor, utilizing the vat polymerization technique. The unit size of each TPMS electrode was unified to 2 mm with a resolution of 200 μm. Consequently, variations in parameters such as surface area, porosity, and density of the TPMS structures were solely attributable to the structure’s category due to their inherent periodicity. This standardization allowed for a systematic evaluation of how TPMS structures impact the electrochemical and mechanical performance of the electrodes and the supercapacitor cells. Subsequently, the 3D-printed precursors underwent a transformation into SiOC ceramic substrates through a sintering process. This process caused preform structure shrinkage due to polymeric precursor pyrolysis (Fig. 1b). Nevertheless, the integrity of the original designs was preserved (Supplementary Fig. 1). The uniform shrinkage of the polymer precursor played a pivotal role in minimizing defect introduction in the resultant SiOC. This outcome was facilitated using highly homogeneous raw materials and a meticulously controlled pyrolysis process, which ensured adequate densification. As a result, the morphology of the SiOC displayed a smooth finish, devoid of visible holes or cracks, effectively reducing stress concentration and enhancing mechanical stability21,24.

A subsequent pyrolytic carbon layer amplified the conductivity of the substrate, realized via the surface coating of polyvinylpyrrolidone followed by a carbonization process, thus creating the C/SiOC current collectors. In situ polymerization of PANI on C/SiOC as the active material resulted in the TPMS structures turning green due to the formation of emeraldine salt PANI (Fig. 1b). This process ensured a uniform and comprehensive coating of PANI directly onto the current collector, establishing a strong and binder-free connection. It not only promoted efficient electron exchange but also minimized the risk of detachment of active materials from substances. Moreover, the approach was simple and nondestructive, preserving the integrity of the original structures. X-ray diffraction (XRD) patterns of SiOC and C/SiOC exhibited a broad peak around 25°, indicative of their amorphous nature21,24. This is attributed to the wide size range of silica nanodomains and the constraints imposed by surrounding mixed bonds and interdomain layers. The XRD pattern of PANI/C/SiOC displayed a small but sharp peak at 25.6°, corresponding to the (200) crystal planes of emeraldine salt (Fig. 1c)32,44. This result confirmed the successful coating of PANI in its emeraldine salt state on the substrate.

PANI/C/SiOC electrodes were subsequently integrated with PVA hydrogel electrolyte utilizing FDM-printed molds to fabricate supercapacitor cells. Two PANI/C/SiOC electrodes, along with a pristine SiOC substrate, sharing the same TPMS structure, were positioned within the mold, serving as the positive and negative electrodes, and the separator, respectively. A mixture of PVA, H2SO4, and borax solution was then injected into the mold to infiltrate the porous electrodes and separator. Subsequent crosslinking through freeze-thaw cycles led to the formation of the PVA hydrogel, which functioned as the electrolyte while concurrently binding the electrodes and separator into a cohesive unit. Post-integration, the assembled cells were detached from the molds, with the electrode architectures were completely preserved within the hydrogel matrix, unaffected by the crosslinking process. This indicated a damage-free and seamless assembly process of electrodes, separators, and electrolytes that can yield uniform and replicable units (Fig. 1d). The resulting cells exhibited impressive lightness, which can be attributed to the low densities of both the electrodes and the electrolyte. The densities for the electrodes G, I-WP, D, and P were measured at 0.76 ± 0.016, 0.96 ± 0.008, 1.03 ± 0.015, and 0.52 ± 0.011 g/cm3, respectively, indicating the lightness and high process stability. The assembled cells exhibited a slight increase in density due the infusion of hydrogel electroltyte, ranging from 1.27 to 1.64 g/cm3 (Fig. 1e). The lightweight can effectively reduce energy self-consumption in practical applications.

The incorporation of the SiOC substrate simultaneously endowed the cell with adaptability to pressure loading. Furthermore, the structural continuity and periodicity between the electrodes and separator contributed to a holistic reinforcement in the cell, resulting in isotropic mechanical behavior. Resultant stress-strain diagrams of the cells exhibited rigid mechanical behavior and interlaminar fracturing post-failure due to the periodic TPMS structures (Fig. 1f)42. While the failure of such rigid structures could potentially generate debris leading to secondary damage, the cohesive nature of the PVA hydrogel effectively retained the fragments, thereby minimizing any secondary damage caused by the shards (Supplementary Fig. 2 and Movie 1). The cells, designed with G, D, P, and I-WP structures, demonstrated compressive stresses of 25.36 ± 1.27, 65.72 ± 2.28, 14.75 ± 1.06, and 70.61 ± 1.53 MPa, respectively, along with corresponding energy absorption capacities of 29.49 ± 1.59, 90.40 ± 2.65, 10.76 ± 1.14, and 92.15 ± 3.78 kJ/m3, indicating the high stability of mechanical behavior and assembly process. (Fig. 1g). Remarkably, an assembly of three Cells I-WP connected in series successfully illuminated an LED under a stress of 10 MPa, showcasing their stress resistance and potential for efficient operation under pressure loading conditions (Fig. 1f).

Characterization of structurally strengthened current collector

The efficient integration of the electrode, separator, and PVA hydrogel electrolyte was also facilitated by the porous TPMS structure of the current collector that promoted the infiltration of the PVA solution into the inner space. Precisely tailored using FDM-printed molds, the assembled cell formed a cohesive unit (Fig. 2a). Besides, the detachable mold, with dimensions meticulously matched to the cell, allowed for complete and nondestructive extraction after the assembly. Furthermore, the C/SiOC TPMS structures can provide adequate conductivity due to the coating of pyrolytic carbon, making it promising as a current collector. X-ray photoelectron spectroscopy (XPS) studied the atomic states of elements present on the SiOC substrate and C/SiOC current collector to evaluate their potential as current collectors. In the XPS survey spectra for SiOC, distinct peaks were observed corresponding to Si 2s, Si 2p, O 1s, and C 1s (Fig. 2b). Notably, the intensities of Si 2s, Si 2p, and O 1s peaks diminished in the C/SiOC spectra, which was attributed to the overlying pyrolytic carbon layer that attenuated the signals of these elements (Supplementary Fig. 3). High-resolution spectra of C 1s further supported the conclusion26,27. The deconvoluted peak observed in the C/SiOC spectrum at 284.8 eV, corresponding to C-C and C=C bonds, showed a relatively higher intensity compared to that from SiOC, signifying the incorporation of a surface carbon layer (Fig. 2c). According to the Raman analysis, the ID/IG of C/SiOC decreased compared to that of pristine SiOC, revealing the lower defect level of the surface carbon sphere layer with lower resistance compared with the SiOC matrix (Fig. 2d)45,46. The coating resulted in an enhancement in the electrical performance of the SiOC substrates. The resistivity and conductivity of pristine SiOC were measured as 6.14 ± 0.28 × 104 Ω mm and 16.3 ± 0.73 × 10−3 S/m, respectively, while those of C/SiOC reaching 10.05 ± 0.45 Ω·mm and 99.62 ± 4.52 S/m, respectively, revealing the potential for current collector application and high stability of the pyrolytic carbon coating process (Fig. 2e).

Fig. 2: Characterization of C/SiOC current collector.
figure 2

a Diagram of the assembly process. XPS spectra of SiOC and C/SiOC: b survey spectra and c deconvoluted peaks of C 1s. d Conductivity of SiOC, C/SiOC, and PANI/C/SiOC. e Raman spectrum of SiOC and C/SiOC. Data were presented as mean ± standard deviation from n = 5 independent samples.

Full size image

Structures and surface morphology of C/SiOC current collector

Scanning electron microscope (SEM) images of the C/SiOC current collector, featuring G, D, P, and I-WP designs, revealed their porous structures (Fig. 3a–d). This porous configuration facilitated electrolyte access to the interior of electrodes, thereby overcoming the limitations of surface-only electrochemical reactions. Consequently, electrodes can be customized into various geometries and dimensions without compromising electrochemical performance. Furthermore, all the structures exhibited a terraced morphology with a resolution of 200 μm, showcasing a precise 3D printing and stable ceramization process conducive to mass production. SEM observation also revealed the successful coating of the pyrolytic carbon layer. The morphology of pristine SiOC presented a smooth and defect-free surface finishing, while that of the C/SiOC exhibited homogeneously distributed carbon spheres that can serve as an efficient electron transport pathway (Fig. 3e, f). These findings further demonstrated that the conductivity of SiOC was effectively enhanced through the employment of the carbon layer coating, thus underscoring its high potential for use in porous current collectors. Additionally, the densities of G, D, P, and I-WP structured C/SiOC were recorded at 0.73, 0.92, 0.99, and 0.50 g/cm3, respectively. Consequently, their material densities (structure density/entity rate) were calculated to be 2.72, 2.64, 2.76, and 2.5 g/cm3, respectively, varied due to the different surface area and coating contents. The maximum value of 2.76 g/cm3 still indicated a marked weight reduction when compared to traditional Ni-based (8.99 g/cm3) porous current collectors, thereby underscoring its superiority in enhancing the gravimetric energy density of the assembled devices.

Fig. 3: Structures and surface morphology of C/SiOC current collector.
figure 3

SEM images of SiOC with a gyroid, b diamond, c primitive, and d I-wrapped package structures. SEM images of e SiOC substrate and f C/SiOC current collector.

Full size image

Electrochemical properties and energy storage mechanism

To assess the energy storage capabilities of the cells, electrochemical evaluations were performed using a two-electrode system. Cyclic voltammetry (CV) analysis, conducted at a scan rate of 10 mV/s, revealed that electrodes with solid plate structure and varying TPMS structures exhibited similar redox peaks (Fig. 4a). This similarity in CV curves, attributable to their identical chemical compositions, corresponded to the redox transitions between the leucoemeraldine base (LB) to the emeraldine base (EB), and the emeraldine base (EB) to the pernigraniline base (PB) of PANI, respectively44,47. Furthermore, the CV curves of the C/SiOC current collector exhibited no redox peaks, indicating that the pseudocapacitive reactions occurred mainly on the PANI active layer (Supplementary Fig. 4). Meanwhile, the consistency of the redox processes in electrodes with different TMPS structures demonstrated little influence from geometries on electrochemical behavior. The robustness of these electrochemical properties was further revealed by CV tests conducted across a range of scan rates (10–100 mV/s). Despite the variation in scan rates, the cells maintained similar and nearly symmetrical CV curves, with identifiable redox peaks even at the elevated rate of 100 mV/s48. This consistency highlighted the high reversibility and stability of the cells under varying scan rates (Supplementary Fig. 5). Additionally, observed peak shift and the increasing separation between peak potentials at higher scan rates can be interpreted as heightened electrochemical polarization correlating with ascendant scan rates. This observation underscores the capacity of the cells to maintain functional integrity under different operational conditions49.

Fig. 4: Electrochemical properties and characterization of PANI/C/SiOC cells.
figure 4

Electrochemical properties of cells with different electrode structures: a CV curves at a scan rate of 10 mV/s, b GCD curves at a current density of 3 mA/cm3, and c Ragone plots (power density vs energy density). d SEM image of PANI/C/SiOC electrodes. e FT-IR spectrum of SiOC, PANI/C/SiOC, and PANI. f XPS deconvoluted peaks of N 1s for PANI/C/SiOC. g FT-IR spectrum of freeze-dried PVA. h SEM image of freeze-dried PVA. i Schematic illustration of the electrochemical reaction process depicting the transitions among different PANI states: leucoemeraldine base (LB), emeraldine base (EB), and pernigraniline base (PB).

Full size image

Galvanostatic charge-discharge (GCD) curves of the cells with different structures, measured at a current density of 3 mA/cm3, exhibited resembled isosceles triangle shapes across electrodes (Fig. 4b), underscoring the reversible nature of the charging and discharging processes50. However, the electrodes with D and I-WP structures exhibited more pronounced charging and discharging plateaus, attributable to more efficient redox processes. Despite each electrode operating at the same volumetric charging and discharging current density, variations in active material loading (stemming from differences in the specific surface areas of each electrode) resulted in differing area-specific charging and discharging current densities. Consequently, the G and P structured electrodes were operated at higher area-specific current densities, leading to less effective engagement of PANI in the redox processes due to inadequate faradaic redox kinetics. Therefore, it can be concluded that while all electrodes can operate at high current densities, those with a higher specific surface area can develop a larger active layer, thereby facilitating more sufficient redox processes and maximizing the electrochemical performance of the active layer. Specific capacitances for the G, D, P, and I-WP structured electrodes were calculated at 516.99, 601.06, 369.45, and 585.51 mF/cm3, respectively, indicating an enhancement when compared to the solid plate structured PANI/C/SiOC electrode (91.77 mF/cm3) and C/SiOC current collector (13.95 mF/cm3, Supplementary Fig. 4 and Table 1). While the TPMS structures do not significantly impact the electrochemical reactions themselves, they do affect specific capacitance by altering surface area availability and the amount of active material loading. GCD curves for each cell across a range of current densities (1–10 mA/cm3) also yielded nearly triangular shapes, revealing the qualified reversibility (Supplementary Fig. 5)25. However, an attenuation in the specific capacitance of each cell was observed with increasing current densities, a likely consequence of electrode resistance and potentially inadequate faradaic redox kinetics at higher currents (Supplementary Table 1). Despite this, the cells demonstrated a wide operational power range while maintaining a notable energy density (Fig. 4c). For example, the I-WP structured electrode achieved an energy density of 97.63 μWh/cm3 at a power density of 0.5 mW/cm3, and maintained an energy density of 62.07 μWh/cm3 even at a higher power density of 5 mW/cm3. Consequently, the corresponding cell demonstrated an energy density of 32.54 μWh/cm3 at 0.17 mW/cm³ and sustained 20.69 μWh/cm3 at 1.7 mW/cm3 (Supplementary Fig. 6). These results underscore the significant potential of the cells for practical applications.

The energy storage capacity of the cells is primarily attributed to the surface PANI layer, which stores and releases electrons via reversible redox reactions. The surface morphology of the PANI/C/SiOC electrodes, characterized by a porous structure with PANI nanorods arrayed homogeneously on the substrate (Fig. 4d), exhibited increased surface roughness and nanopores compared to the smooth C/SiOC current collector (Fig. 3f). This porous morphology led to enlarged accessible areas of the electrode, enhancing electrolyte penetration and ion transfer efficiency44. The nanorod morphology arose from the formation of amphiphilic aniline sulfate salt, which self-assembled into micelles in aqueous solution, serving as soft templates for the in situ arraying of PANI nanorods on the C/SiOC surface42. Besides, the in situ polymerized PANI further decreased the resistivity of the electrodes, attributable to the inherent high conductivity of PANI (Fig. 4e). Fourier-transform infrared spectroscopy (FT-IR) characterization of the active material confirmed the characteristic peaks of PANI (Fig. 3e)47,50, including the C = C stretching mode of quinonoid rings at ~1559 cm−1, benzenoid units around 1480 cm−1, C-N stretching mode at 1301 and 1244 cm−1 characteristic of aromatic amines, the C-H out-of-plane bending vibration near 800 cm−1 of 1,4-disubstituted rings, and the N-H stretching vibration around 3300 cm−1. XPS analysis further studied the atom states on the electrode surface (Fig. 4f and Supplementary Fig. 7). High-resolution N 1 s spectra revealed the presence of quinoid imine (-N=), benzenoid amine (-NH-), and positively charged nitrogen (N+) species within PANI. The occurrence of N+ ions indicated PANI in its conductive, doped state, consistent with its emeraldine salt (ES) configuration51,52,53. These analytical results affirm that the in situ polymerization process preserved the chemical composition and microstructural integrity of PANI, thereby ensuring its effective electrochemical functionality.

The qualified electrochemical properties of the cells were also attributed to the effective performance of the PVA hydrogel electrolyte, which was filled with the electrolyte solution to ensure sufficient contact with the active materials. The electrolyte retention capability of the PVA hydrogel is attributed to the abundance of hydroxyl groups within its polymer structure, facilitating hydrogen bond formation with H2O molecules. FT-IR spectroscopy of the PVA revealed a broad absorption band between 3100–3700 cm−1, indicative of dissociated O-H groups in the PVA chains (Fig. 4g)54. Additional bands around 2916 and 1431 cm−1 were attributed to aliphatic C-H stretching and -CH2 bending vibrations, respectively55. The peak at 1094 cm−1 corresponded to the stretching vibration of the secondary alcohol C-O, signifying the intermolecular connections within PVA. Crucially, the 1660 cm−1 peak evidenced the C=O stretching vibration of the boric ester bond, confirming the crosslinked structure of PVA55. Morphologically, PVA exhibited an interconnected porous framework, enhancing the absorption of the electrolyte solution, thereby ensuring effective electrolyte interaction with the electrode surfaces (Fig. 4h). Additionally, the PVA hydrogel between the electrodes maintained an ionic conductivity of 16.29 S/m (Supplementary Fig. 8). Thus, the energy storage mechanism in these cells can be elucidated by the synergistic interaction between the conductive PANI/C/SiOC electrodes and the hydrogel electrolytes, promoting substantial ionic exchange between the PANI and the electrolyte. Furthermore, the pyrolytic carbon-coated SiOC served electron transit during charge/discharge processes, and PANI acted as an electron reservoir, storing and releasing electrons by redox transitions between its various states (Fig. 4i).

Electrochemical stability post-impact and under-loading

Despite demonstrating satisfactory static electrochemical performance, the capability of cells to sustain normal function under rigorous conditions, such as post-impact scenarios, complex loading, and in the presence of electrolyte damage, remains a critical consideration. To assess the impact resistance of the cell, it was subjected to varying impact loads oriented both parallel and perpendicular to the electrode (Fig. 5a). The specialized experimental setup tailored for these small-scale cells is depicted in Fig. 5b, with the I-WP cell selected for testing owing to its optimized energy absorption properties. Impact energy was varied by dropping a 200 g weight from different heights (Supplementary Fig. 9 and Movie 2). It was observed that the cell remained intact and visibly undamaged at impact energies from both directions up to 0.3 J/cm3. However, at an energy level of 0.4 J/cm3, damage to the internal SiOC electrode was evident, with noticeable edge breakage. Consequently, the electrochemical performance of the Cell I-WP was evaluated post-impact at an energy level of 0.3 J/cm3. CV tests over a range of 10–100 mV/s revealed that the post-impact CV curves closely resembled those obtained under static conditions, with characteristic peaks retained (Fig. 5c). This consistency suggested the preservation of electrochemical reaction capability after impact. Additionally, GCD tests confirmed charge/discharge reversibility, with the curves maintaining their quasi-triangular shapes (Fig. 5d). Specific capacitances of the I-WP electrode post-impact were measured at 699.83, 640.03, 582.54, 537.27, 478.85, and 430.18 mF/cm3 at current densities of 1, 2, 3, 5, 8, and 10 mA/cm3, respectively, demonstrating negligible variation from the static test results (Supplementary Table 2). The electrochemical impedance spectroscopy (EIS) also displayed minimal deviation from the static tests, indicating a solution resistance of 11.74 Ω (Supplementary Fig. 10a). This consistent electrochemical behavior suggests that the electrolyte was also well protected since the majority of the impact energy was effectively absorbed by the SiOC strengthened electrodes.

Fig. 5: Electrochemical performance post-impact and under dynamic pressure loading.
figure 5

a Schematic illustration of the impact process. b Impact energy applied on the Cell I-WP. c CV curves of Cell I-WP at different scan rates after the impact of 0.3 J/cm3. d GCD curves of Cell I-WP at different current densities after impact of 0.3 J/cm3. e Compressive stress curves of the cell showing the dynamic loading process. f Illustration of the electrochemical testing under dynamic pressure loading. g CV curves of Cell I-WP at different scan rates under dynamic pressure loading. h GCD curves of Cell I-WP at different current densities under dynamic pressure loading. i Compressive stress curves of SiOC substrates with TPMS structures. j High-resolution transmission electron microscope (HRTEM) image and diffraction pattern of SiOC. k Simulation showing the compression strain distribution of the TMPS structures.

Full size image

The second evaluation emulated dynamic pressure loading to approach practical application conditions. A dynamic strain from 0 to 0.5%, at a rate of 0.01%/s, was applied to the Cell I-WP, generating a stress range from 0 to 18.8 MPa (Fig. 5e and Supplementary Movie 3). Concurrently, the electrochemical properties were monitored under these cyclic strains, with the cell connected to an electrochemical workstation for real-time data capture (Fig. 5f). CV curves under dynamic loading mirrored the electrochemical performance from static tests, with identifiable reversible redox peaks and a linear increase in CV curve area with scan rate, indicating consistent energy storage and release processes (Fig. 5g). GCD curves corroborated this reversibility, exhibiting quasi-triangular shapes (Fig. 5h). Additionally, both the specific capacitance and EIS spectrum showed negligible variation compared to those under static state, confirming the reliable functionality of the cell under dynamic loading (Fig. 5h and Supplementary 10b).

The load-bearing capacity of the cells predominantly arose from the robust ceramic-based current collector. The SiOC substrate with G, D, P, and I-WP structures exhibited compressive strength of 26.63 ± 1.19, 72.39 ± 2.53, 13.94 ± 0.95, and 75.85 ± 2.04 MPa, respectively, and corresponding Young’s modulus of 1.34 ± 0.09, 2.81 ± 0.12, 1.10 ± 0.06, and 2.71 ± 0.11 GPa (Fig. 5i). In contrast, the compressive stress curve of solid SiOC indicated multiple peaks indicative of structural flaws and registered a compressive strength of 46.79 ± 5.6 MPa lower than I-WP structured SiOC, with a higher standard deviation due to the defects (Supplementary Fig. 11). It is important to note that the creation of hollow structures within SiOC ceramics can effectively mitigate the formation of cracks and pores resulting from the pyrolysis of bulk polymeric precursors18,20. Thus, I-WP structured SiOC demonstrated a more uniform stress distribution with enhanced compressive strength, revealing that this configuration not only reduced weight and increased surface area, but also optimized mechanical behavior. Additionally, the specific compressive stress (SCE) and specific modulus (SM) of the C/SiOC current collectors were calculated to assess the structural reinforcement capabilities of various TPMS structures (Supplementary Fig. 12a and Eqs. (5)–(6)). The C/SiOC current collectors featuring G, D, P, and I-WP configurations registered SCE of 99.96 ± 4.43, 208.41 ± 7.26, 70.96 ± 4.76, and 212.42 ± 5.71 MPa, respectively. Notably, the I-WP structured C/SiOC demonstrated the highest specific compressive stress, indicating efficient structural reinforcement compared to other designs with the same cell size and resolution.

Furthermore, flexural stresses for these structures were measured at 42.62, 52.78, 5.94, and 61.83 MPa, respectively (Supplementary Fig. 12b, c), indicating robust bending resistance. The robustness was primarily derived from the amorphous nature of SiOC, which eliminated grain boundaries, leading to a uniform stress distribution (Fig. 5j and Supplementary Fig. 13)21. Moreover, the defect-free finish of the SiOC surface was devoid of discernible holes or defects (Fig. 3e). However, though high compressive resilience of the SiOC substrates was maintained in the corresponding cells, a slight decrease in compressive stress was noted (Fig. 1f, g), possibly due to defects induced by volume expansion of hydrogel during the freeze-thaw processes. Especially for Cell D, which exhibited a more significant decline in mechanical performance due to its high structural density that hindered the uniform volatilization of small molecules during pyrolysis, inducing defects and inconsistent shrinkage. The stable performance of the Cell I-WP after impact and under dynamic pressure loading was also underpinned by the even deformation of its TPMS structure. Simulations revealed a homogeneous strain distribution in TPMS structures, with the D and I-WP structures exhibiting a more uniform stress distribution, mitigating localized stress concentrations (Fig. 5k). This uniformity also ensured continuous connectivity between C/SiOC and PANI, preventing separation. Additionally, the inherent flexibility and adhesion of the PVA hydrogel maintained a seamless interface between electrodes and electrolytes, enhancing the ability of the cells to function under complex loading conditions40.

Repairability of PVA electrolyte and scalability of the cell

Still, the PVA hydrogel electrolyte remained susceptible to damage, as it was not entirely encapsulated by the rigid electrodes. Fortunately, the hydrogel electrolyte exhibited self-healing properties, a feature ascribed to its hydrogen and boric ester bonds56, potentially reducing its vulnerability to damage. To simulate extreme damage scenarios for evaluating self-healing performance, the Cell I-WP was artificially bisected along the separator and electrode interface, cutting only through the hydrogel. The two separated halves were then manually realigned and let stand for 30 s. Then, the separated cell can be restored to its original state effectively (Fig. 6a). To evaluate the influence of this self-healing on the electrochemical behavior of the cell, the electrochemical properties of the restored Cell I-WP was assessed. CV analyses revealed that the repaired cell retained electrochemical reaction capacity comparable to that of the undamaged cell, consistently revealing the redox processes (Fig. 6b). Concurrently, GCD curves showed similar charge/discharge shapes for cells both pre and post self-healing (Fig. 6c). A slight reduction in specific capacitances was observed, potentially resulting from the partial loss of electrolyte solution during the separation, which increased the internal resistance of the hydrogel electrolyte (Supplementary Fig. 14 and Table 2). Nonetheless, most of the initial capacitance was retained, indicating an effective repair of the ion exchange pathway. The self-healing characteristic of the PVA hydrogel is attributable to the reversible nature of both hydrogen and boric ester bonds55,56. Upon recontact, these cleaved bonds underwent reformation, facilitating the regeneration and self-repair of the crosslinked network (Figs. 4g, 6d). This self-healing ability significantly enhances the resistance to electrolyte perturbations, thereby increasing its threshold for damage tolerance. It also ensured high working stability under more complex and practical conditions. Electrochemical behavior and capacitance of Cell I-WP remained largely unchanged after 100 cycles of twisting and stretching at the interfaces between electrodes and separators, respectively, demonstrating significant adaptability to damaging environmental conditions (Supplementary Fig. 15 and Table 2).

Fig. 6: Self-healing performance of PVA hydrogel and scalability of the cells.
figure 6

a Schematic illustration showing the self-healing process. b CV curves of self-healed Cell I-WP at different scan rates. c GCD curves of self-healed Cell I-WP at different current densities. d Schematic illustration showing the possible self-healing mechanism. e Cycling stability of self-healed Cell I-WP. f XRD pattern and g XPS C 1s deconvoluted peaks of PANI/C/SiOC after cycling test. h Ragone plots (power density vs energy density) of Electrode I-WP post-impact, under dynamic loading, and after self-healing. Photography showing the i interlocking structure and j scalability of the cell.

Full size image

The long-term operational stability of the self-healed cell is a critical factor for its practical application. Accordingly, we assessed its cyclic durability for 5000 cycles at a current density of 10 mA/cm3. The cell demonstrated a rapid capacitance attenuation to 83.86% of its initial value after 1000 cycles, while remaining relatively stable during the rest cycles that maintained 80.31% capacitance, reflecting an acceptable level of cyclic stability and reliability of the repaired PVA hydrogel electrolyte (Fig. 6e). Moreover, the cell displayed higher capacitance retention compared to pristine PANI50, due to the in situ polymerization process that effectively anchored PANI nanorods onto the current collector, mitigating detachment risks57. SEM images of the electrodes after 1000 cycles revealed alterations in the nanorod morphology, possibly leading to a reduced surface area for electrolyte interaction and resulting in decreased capacitance (Fig. 6e). The XRD pattern of the PANI/C/SiOC electrode after cycling tests consistently exhibited the peak at 25.6°, which corresponded to the (200) crystal planes of PANI (Fig. 6f). Concurrently, XPS analysis indicated minimal changes in the atomic states, except that the high-resolution N 1s spectra showed a decrease in the intensity of the benzenoid amine -NH-, attributed to the formation of pernigraniline base following complete discharge. Furthermore, high-resolution Si 2p spectra revealed no significant alterations from its initial state, suggesting that the C/SiOC current collector did not participate in the pseudocapacitive reaction (Fig. 6g and Supplementary Fig. 16). Therefore, it is inferred that the observed reduction in capacitance after cyclic tests primarily arises from the deterioration of the surface PANI layer, while the self-healed electrolyte maintains its functional integrity.

Furthermore, the energy and power densities of the Electrode I-WP under various extreme conditions were compared (Fig. 6h). The self-healed electrode I-WP demonstrated a power range extending from 0.5 to 5 mW/cm3, with corresponding energy densities varying between 95.50 and 9.58 μWh/cm3, indicating little decline compared to those of undamaged cell. The minimal variation in electrochemical performance observed among the static state, post-impact state, dynamically loaded state, and self-healed state underscored the remarkable resistance to extreme environments. Additionally, Supplementary Fig. 17 details the practical energy and power densities of the Cell I-WP under different scenarios. This analysis, diverging from traditional approaches that primarily focus on the performance of active materials, offered a realistic assessment of the device’s functional capabilities under actual operating conditions. Currently, efforts to develop 3D-printed PANI-based symmetric supercapacitors have gradually enriched the existing body of research (Supplementary Table 3). However, their performance under complex conditions has rarely been explored58,59,60,61,62. The PANI/C/SiOC electrode not only demonstrated comparable capacitance to the 3D-printed PANI-based symmetric supercapacitors but also excelled due to its ability to operate under harsh conditions. In comparison with other rigid supercapacitor electrodes or cells, the Cell I-WP showed enhanced compressive strength and energy absorption, managing more practical yet extreme conditions with high damage invulnerability38,63,64,65,66,67. Future modifications of the PANI/C/SiOC supercapacitor should aim to enhance temperature adaptability, such as frost and drying resistance, thereby extending its application to more severe environments.

Another advantage of 3D printing lies in its flexibility in cell architecture design. Although this work employed a conventional electrode shape to standardize the calculation and evaluation of theoretical performance, the further design of seamless joints between electrodes and separators is straightforward and adaptable. The design of interlocking structures can offer a diverse assembly approach (as illustrated in Fig. 6i), enabling the creation of customized and stable devices tailored to various application requirements with firmer and more robust assembly manners. The potential for scalable production, afforded by the vat polymerization technique, further highlights the practical applicability of this fabrication approach (Fig. 6j). This renders the supercapacitor cells versatile, and adaptable across various geometries for divergent applications. For instance, cells can be fabricated to dimensions akin to smartphones or 18650 batteries with diameters and heights of 18 and 65 mm, respectively. For the 18650-sized cell in Fig. 5j, the anode can be printed into a cylinder with a diameter of 5.3 mm, and the cathode was a hollow cylinder with a thickness of 1.7 mm to ensure symmetry, while the thickness of the hydrogel was 2 mm. Notably, the cell design can also incorporate connectors, enabling series or parallel configurations. The incorporation of bespoke connectors and sockets into the cells can accommodate various assembly methods, streamlining traditional splicing and assembly processes that typically involve complex circuits and wires. This versatility is instrumental in the assembly of expansive supercapacitor devices or modulating output voltages to cater to a spectrum of applications. Furthermore, this structurally strengthened strategy can be promoted to other supercapacitor material systems, since the active materials coated on the current collector and electrolyte solution swollen in the hydrogel are replaceable. Computer-aided manufacturing also supports real-time adjustments to processes and standardization of production. Furthermore, the ease of sharing digital structures could enhance the establishment of production standards and foster collaborative development68.

Discussion

The current study reports the strategy for the development of ready-to-use supercapacitor cells characterized by their impact-resistant, load-bearing, and self-healing capabilities. Utilizing vat polymerization, SiOC was 3D printed into four distinct TPMS structures, including G, D, P, and I-WP. Subsequent coating with pyrolytic carbon enhanced conductivity to 99.62 S/m. PANI was in situ polymerized onto the C/SiOC current collectors, forming nanorod-arrayed microstructures that effectively increased the exposure of active areas. These lattice electrodes, when integrated with SiOC separators and infused by PVA hydrogel electrolyte, formed usable cells with tunable capacitance and mechanical properties based on the adjustment of the TPMS structures. The Electrode I-WP, with optimized mechanical performance and largest surface area, showcased a specific capacitance of 585.51 mF/cm3 at 3 mA/cm3. Its energy density peaked at 97.63 μWh/cm3 with a power density of 0.5 mW/cm3 and retained 62.07 μWh/cm3 at 5 mW/cm3. Mechanical evaluations of the Cell I-WP revealed compressive stress, Young’s modulus, and energy absorption of 70.61 MPa, 2.75 GPa, and 92.15 kJ/m3, respectively. With a density of 1.6 g/cm3, the Cell I-WP stood out as both lightweight and robust. The study further investigated the electrochemical performance of the Cell I-WP under three extreme conditions: post-impact, under dynamic pressure, and after self-healing. The cell maintained normal operation after enduring impact energy of 0.3 J/cm3, dynamic loading varying from 0 to 18.83 MPa, and self-healing from bifurcation, preserving most of its original capacitance. This electrochemical stability was attributed to the compression resistance of the electrodes, preserving the integrity and connectivity of active materials. Concurrently, the flexibility of hydrogel ensured consistent electrolyte-electrode connectivity, even during deformation. Besides, the hydrogen and boric ester bonds in PVA chains provided the cells with self-healing capability. The energy and power densities of the Cell I-WP varied from 20.69 to 32.54 μWh/cm3 and 0.17 to 1.67 mW/cm3, respectively, demonstrating the practical performance. The design flexibility afforded by vat polymerization further allowed for the incorporation of interlocking and bespoke connection structures on the cells, enhancing their adaptability for assembly. The potential for mass production and the versatility of SiOC-based electrode structures further enhance the scalability of this approach. Overall, this research highlights the feasibility of enhancing supercapacitor safety using mechanically robust electrodes and self-healable electrolytes, offering a strategy for the design of supercapacitors for harsh environments that can be promoted in other supercapacitor systems.

Methods

Structure design of electrodes

Figure 1a depicts the configuration of polyvinyl alcohol hydrogel-infilled PANI/C/SiOC supercapacitor cell reinforced by C/SiOC current collectors with different triply periodic minimal surface (TPMS) structures. Four TPMS designs, namely Gyroid, Diamond, Primitive, and I-wrapped package (G, D, P, and I-WP, with mathematical expressions provided in the Supplementary Information Eqs. (1–4)), were created by Matlab software and converted into STL-format models for import into a 3D printer to produce preforms. The resolution was set at 200 μm to balance the inherent accuracy limitations of 3D printers with the desired increase in surface area and reduction in weight of the fabricated structures. The unit size of each TPMS structure was unified at 2 mm. The porosity of the G, D, P, and I-WP structures were 73.2, 65.2, 80.1, and 64.2%, respectively, with a specific surface area of 31.35, 38.92, 24.02, and 34.95 cm2/cm3.

3D printing of C/SiOC current collectors

The TPMS-structured current collectors were fabricated using a digital light processing (DLP) printer (Max UV385, Asiga) equipped with a 385 nm UV light source. The substrates were 3D printed from a photocurable-modified methyl-silsesquioxane resin (SILRES MK, Wacker-Chemie). During the printing process, the exposure time of each layer was maintained at 2 s, the current intensity was set at 15 mW/cm3, and the layer thickness was 50 μm. The photocurable resin feedstock was synthesized by dissolving 50 g MK resin and 25 mL 3-methacryloxypropyl trimethoxy (KH570, Shanghai yuanye Bio-Technology Co., Ltd China) in an acidic solvent (by adding 15 μL of standard aqueous hydrochloric acid solution) consisting of 15 mL tetrahydrofuran (THF, Sigma-Aldrich, Singapore) and 15 mL tripropylene glycol monomethyl ether (TPM, Macklin, China). For enhanced perform strength and UV polymerization initiation during the VP printing, 1.5 g phenyl bis (2, 4, 6-trimethyl benzoyl) phosphine oxide (photoinitiator 819, FCOM Co. Ltd. China) and 40 mL trimethylolpropane triacrylate (TMPTA, FCOM Co. Ltd. China) were incorporated to the above solution for completing the feedstock synthesis. Owing to the PDC’s shrinkage after debinding, the digital model was proportionally enlarged (160%) prior to printing based on the expected shrinkage rate. This adjustment ensured that the resultant components retained the designed resolution post-sintering. After printing, the preforms were pyrolyzed in argon at 900 °C (with a heating rate of 1 °C/min) for 2 h for debinding, yielding dense SiOC ceramic structures. The SiOC was subsequently dip-coated using polyvinylpyrrolidone (PVP, Sigma-Aldrich, Singapore) solution that was prepared by dissolving 0.3 g PVP into 10 mL ethanol (Sigma-Aldrich, Singapore). The coated structures were sintered at 1000 °C (with a heating rate of 5 °C/min) for 5 h in argon for the carbonization of PVP coating, resulting in the formation of a conductive pyrolytic carbon layer, completing the C/SiOC current collector.

Fabrication of PANI/C/SiOC electrode

PANI was then in situ polymerized onto the current collector. Aniline (Sigma-Aldrich, Singapore) was dissolved in a 0.5 M H2SO4 solution to yield a 0.05 M solution. The C/SiOC current collectors were immersed in this solution for adequate infiltration. Upon the addition of a 0.05 M ammonium persulfate (APS, Sigma-Aldrich, Singapore) solution with the same volume as the PANI solution, the polymerization was initiated. Conducted in an ice bath for 8 h, this process yielded PANI/C/SiOC electrodes.

Assemble of hydrogel electrolyte infused PANI/C/SiOC supercapacitor

PANI/C/SiOC electrodes were integrated with polyvinyl alcohol (PVA, Sigma-Aldrich, Singapore) hydrogel, dissolving 1 M H2SO4 solution, to assemble a ready-to-use supercapacitor cell. Briefly, 2.00 g PVA was dissolved in 10 mL of the 1 M H2SO4 solution at 85 °C. Upon cooling, a 0.04 mol concentration of 10 mL Na2B4O5(OH)4·8H2O (borax, Sigma-Aldrich, Singapore) was added and mixed thoroughly. This mixture was injected into a composite structure consisting of two PANI/C/SiOC electrodes and a pristine SiOC separator that shared the same structure, and was shaped by fused deposition modeling (FDM, Original Prusa i3 MK3S) printed polylactic acid (PLA, Prusa 3D) molds into the hydrogel-infused lattice configuration (Fig. 2a). The PVA inside the molds was the crosslinked by a freeze-thawing process to create a hydrogel electrolyte within the two electrodes and separator, yielding symmetrical supercapacitor cells in a sandwich configuration. The cells with the Gyroid, Diamond, Primitive, and I-wrapped package structured electrodes were denoted as Cell G, Cell D, Cell P, and Cell I-WP, respectively. For comparison, PANI/C/SiOC electrodes and C/SiOC current collector without TPMS structure were also assembled into cells using the above method, which denoted as Cell plate and Cell C/SiOC, respectively.

Characterization

Electrode morphology was investigated using scanning electron microscopy (SEM, Hitachi S-4700), while the microstructure of SiOC ceramic was studied via transmission electron microscopy (TEM, FEI Talos F200X). X-ray diffraction (XRD, Bruker D8 Advance) patterns were obtained using Cu Kα radiation (λ = 0.15418 nm) over a 2θ range of 10 to 80°. Raman spectrometer (Renishaw inVia) studies the carbon microstructure. Surface chemical compositions were studied through X-ray photoelectron spectroscopy (XPS, Thermo Fisher K-Alpha) with Al-Kα irradiation, referencing the C 1s peak at 284.8 eV for binding energy (BE). Chemical structures were assessed using Fourier-transform infrared spectroscopy (FT-IR, Nicolet iS5, Thermo Fisher, America) across a wavenumber range of 400–4000 cm−1. The average density was measured by dividing mass by volume using 5 different samples, and the mean value, along with the standard deviation, was reported.

Electrical conductivity

The conductivity of the structures was assessed using a standard four-probe method, employing a physical property measurement system (KeithLink four-point conductivity probe measurements). Samples were fabricated into rectangles measuring 2.0 × 0.5 × 0.1 cm with a fully dense configuration. During measurements, the four-pin probe was in firm contact with the surface of the structure while resistance values were recorded. Electrical conductivity was calculated using the equation:

$$\sigma=1/\rho=1/({R}_{{{{\rm{s}}}}}\times {T}_{{{{\rm{f}}}}})$$
(1)

Where σ (S/mm), ρ (Ω • mm), Rs (Ω), and Tf (mm) represent electrical conductivity, resistivity, resistance, probe spacing, and thickness, respectively. Each measurement was taken on 5 different samples, and the mean value, along with the standard deviation, was reported.

Electrochemical test

The electrochemical behavior of the assembled supercapacitor cells was evaluated using a two-electrode configuration on an electrochemical workstation (VMP3, Biologic Inc.). A connection structure was designed on each current collector to connect to the workstation (Fig. 4a). The connection structure was coated by a pyrolytic carbon layer along with the current collector to ensure a minimized internal resistance when accessing the workstation, but without being modified by PANI and filled with PVA hydrogel. Cyclic voltammetry (CV) analyses were performed within a 0–1 V potential window, with scan rates spanning 10–100 mV/s. Galvanostatic charge-discharge (GCD) tests were conducted over a 0–1 V potential range at current densities between 1–10 mA/cm3. Electrochemical impedance spectroscopy (EIS) measurements ranged from 200 kHz to 0.01 Hz, set at 0 mV relative to open circuit potential. Dimensions of the electrodes and separator were both 10 × 10 × 2 mm, assembling a cell unit with the dimensions of 10 × 10 × 6 mm. Volume-specific capacitance of the electrode (Cv, mF/cm3) was derived from GCD curves:

$${C}_{{{{\rm{v}}}}}=I\triangle t/v\triangle V$$
(2)

where I (mA) represents the current, ∆t (s) represents the discharge time, and ∆V (V) represents the potential window during the GCD process, v (cm3) represents the volume of the electrode. Energy density (E, mWh/cm3) and Power density (P, mW/cm3) of the electrodes were defined as:

$$E=({C}_{{{{\rm{v}}}}}{\triangle V}^{2})/(2\times 3.6)$$
(3)
$$P=3600\times E/\triangle t$$
(4)

Volume-specific capacitance (Ccell) of the assembled cell was calculated to evaluate the practical performance of the devices using:

$${C}_{{{{\rm{cell}}}}}=I\triangle t/{v}_{{{{\rm{cell}}}}}\triangle V$$
(5)

where I (mA) represents the volume of the cell, vcell (cm3) represents the volume of the electrode. The energy density (Ecell, mWh/cm3) and Power density (Pcell, mW/cm3) of the assembled cells were calculated using:

$${E}_{{{{\rm{cell}}}}}=({C}_{{{{\rm{cell}}}}}{\triangle V}^{2})/(2\times 3.6)$$
(6)
$${P}_{{{{\rm{cell}}}}}=3600\times E/\triangle t$$
(7)

The post-impact electrochemical performance was assessed following the application of varying impact energies to the cell. This energy was generated by releasing a 200 g weight, allowing it to free fall from diverse heights. To uniformly distribute the impact energy, two plexiglass boards were placed above and below the cell. The specific impact energy (Ei, in joules per cubic centimeter, J/cm3) was calculated using the formula:

$${E}_{{{{\rm{i}}}}}={mgh}/{v}_{{{{\rm{cell}}}}}$$
(8)

Where m is the mass of the weight (200 g), g represents the acceleration due to gravity, h (m) denotes the height of the fall.

Dynamic pressure-loaded electrochemical performance was gauged by subjecting cells to strains from 0 to 5% using a universal testing machine (Instron 5848 micro-tester), concurrently with electrochemical measurements. The loading rate stood at 0.01 mm/s, generating stresses between 0 and 18.8 MPa. The cells were simultaneously connected to the electrochemical workstation, and their electrochemical performance was recorded under such loading to evaluate stability under complex conditions. To study practical performance as a power supplier, three fully charged Cell I-WP units were connected in series and subjected to a stress of 10 MPa using the universal testing machine. The cells were then connected to an LED light under this stress to assess their functionality under pressure loading.

Ionic conductivity (σ, S/m) of the PVA hydrogel was gauged with carbon paper as current collectors to replace electrodes, maintaining the same fabrication and measurement conditions as the cells. The σ was calculated by:

$$\sigma=d/({R}_{{{{\rm{s}}}}}\times A)$$
(9)

where d (m) denotes the thickness of the hydrogel electrolyte between the current collectors, Rs (Ω) represents the ohmic resistance obtained from electrochemical impedance spectroscopy, A (m2) is the interface area of the hydrogel with the collectors.

Mechanical test

Compressive performance of the cells and SiOC substrates were assessed using an electronic universal testing machine (shimadzu AG25TB Testing Machine) at a loading rate of 0.1 mm/min. The cell was tested in its original state with the dimension of 10 × 10 × 6 mm to evaluate the practical performance, while the SiOC substrate was printed into cubic with the dimension of 10 × 10 × 10 mm to evaluate the theoretical performance. Compressive strength (σc, MPa) was defined by:

$${\sigma }_{{{{\rm{c}}}}}={F}_{\max }/{S}_{\min }$$
(10)

Herein, Fmax (N) and Smin (m2) denote the maximum applied force and minimum cross-section area of the sample, respectively. Young’s modulus (GPa) was obtained from the slope of the stress-strain curve, and the energy absorption (kJ/m3) was calculated by the integral of this curve up to peak stress. Compressive deformation behavior was simulated using COMSOL Multiphysics software. Three-point bending tests were performed using a three-point bending test fixture. Maximum flexural stress (σf, MPa) was calculated as:

$${\sigma }_{{{{\rm{f}}}}}=3(F\times a)/(b\times {d}^{2})$$
(11)

where F denotes the fracture force, a (mm) is the support span, b (mm) and d (mm) denote the width and thickness of the sample, respectively. The flexural modulus (Ef, GPa) is expressed as:

$${E}_{{{{\rm{f}}}}}=({a}^{3}\times m)/4(b\times {d}^{3})$$
(12)

where m is the slope of the load-displacement curve. The support span of the fixture is 15 mm, and the specimens had dimensions of 4 mm in width and 3 mm in thickness. Each measurement was taken on five different samples, and the mean value, along with the standard deviation, was reported.