Natural‐Inspired Ultra‐Flexible Lithium‐Ion Batteries With Fractal Fern Architecture for Extreme Bending and Long‐Lived Stability

ABSTRACT

Flexible lithium‐ion batteries (FLIBs) have attracted great attention as the most promising power sources for wearable electronic devices. However, achieving tolerance to dynamic bending at small radius, adaptability under dynamic loads, and stable electrochemical performance are still key challenges in the design of flexible LIBs. Herein, inspired by the exceptional flexibility of fern plants during expansion and contraction, a novel segmented node deformation LIB with ultra‐flexibility and cycle stability is designed and fabricated. With the unique features, the battery can even withstand severe dynamic alternating bending stress with a radius as small as 10 mm, demonstrating excellent resistance to extreme bending. Combining finite element analysis (FEA) with experimental results, the concentrated stress is effectively dispersed via controllable nodal deformation. The battery thus exhibits ultra‐long cycle stability even under harsh conditions with a minimal bending radius of 10 mm and over 42 000 dynamic bending cycles. Furthermore, the fern‐like battery is seamlessly integrated into both the flexible LED screen and the worm‐like robot. Beyond providing stable power, it functioned as a load‐bearing, deformable component. This work pioneers a novel biomimetic approach to designing next‐generation flexible batteries that simultaneously achieve exceptional mechanical robustness and multifunctionality.

A biomimetic fern‐like lithium‐ion battery utilizes a segmented node design to achieve extreme bending tolerance. Manufactured via scalable punching technology, this structure safely distributes stress during deformation. The resulting batteries showcase ultra‐long cycle stability and integrated load‐bearing capabilities, offering a compelling, scalable solution for demanding, flexible, and wearable electronic applications.

1. Introduction

Flexible electronic devices have been widely applied, such as wearable electronics [1, 2, 3], flexible displays [4, 5], and the medical & health fields [6, 7, 8]. Lithium‐ion batteries (LIBs) are regarded as the primary power source for flexible devices owing to high energy density, mature manufacturing process, and compatibility with electronic equipment [9, 10, 11]. However, the inherent rigidity of LIBs proves fundamentally incompatible with the flexibility requirements of next‐generation devices, precluding reliable operation under harsh mechanical deformations characteristic of flexible system operation [12]. As a result, the novel design of FLIBs lies in ensuring that the mechanical properties of flexible devices are met without compromising electrochemical performance [13, 14, 15].

Currently, substantial research is focused on engineering flexible configurations of FLIBs [16, 17, 18, 19, 20]. The structural design has been a superior solution for balancing flexibility and electrochemical performance. It involves reengineering conventional battery materials into a specific structure, enabling LIBs to simultaneously achieve mechanical adaptability and retain high energy density, such as spine‐like battery, human joint‐inspired battery, scale‐inspired overlapping battery, and DNA‐inspired spiral battery [21, 22, 23, 24, 25, 26]. The flexible battery designed with the above structure achieves functional decoupling through partitioning into energy storage and deformation units. Although mechanical adaptability and electrochemical integrity can be controlled by periodic thickness variations within the battery during operational deformations, unfortunately, a highly complex manufacturing process is incompatible with the existing lithium battery production lines. In addition, the integration will inevitably lead to increased thickness, thereby compromising the original portability and comfort of products such as wearable devices and ultrathin flexible displays.

Ferns are a common plant, and their leaves exhibit a distinctive hierarchical structure, enduring extreme curling and repeated bending without structural compromise due to their ability to deform locally. Inspired by this mechanism, an ultra‐flexible FLIBs with a fern‐like structure is designed to achieve extreme bending and long‐lived cycle stability. The main stem and feathery leaves of a fern are abstracted as a series of hinge‐like connecting components and a horizontal fractal pattern, respectively. Under this design, the electrode is punched into a fern‐shaped structure, creating discrete energy storage blocks (representing leaves) that are interconnected by curved grooves. This configuration provides mechanical flexibility and ensures stable electrical contact via the continuous conductive pathways flanking the grooves (representing the main stem). Importantly, the fern‐like structure effectively isolates the energy storage blocks from high‐stress regions, preventing structural failure. Impressively, this design only requires corresponding blanking of the electrodes rather than complex manufacturing processes, which can maximize the use of the existing full‐cell scale production lines, offering significant commercial potential.

Herein, the fern‐like FLIBs demonstrate ultrahigh flexibility and stable electrochemical performance, retaining 95.7% capacity after > 42 000 dynamic bending cycles at an extreme curvature (bending radius: 10 mm; battery length: 106 mm). More importantly, its capacity can reach 83.4% of that of conventional stacked batteries, and geometric customization can be enabled for diverse operational demands. The mechanistic basis for the synergistic combination of mechanical resilience and cycle stability is systematically revealed by the combination of theoretical derivation, finite element analysis, and experimental results. Under nominal bending conditions (R = 33 mm), conventional stacked batteries exhibit a maximum principal stress of 136 MPa that exceeds the yield strength of Al foil, whereas the fern‐like architecture effectively distributes the mechanical load and reduces the maximum stress at the nodes to only 24.8 MPa. Even under an extreme bending radius of 10 mm, the maximum stress of the proposed battery remains at 52 MPa, still within the safe elastic range. Furthermore, when integrated into flexible LED displays and soft robots, the battery demonstrates multifunctional capabilities such as high energy storage, excellent structural support, and extreme flexibility. It is believed that the unique fern‐like deformation strategy, together with the verified theoretical model, will provide new insight for the design of biomimetic FLIB with excellent electrochemical and mechanical stability, and further provide a potential commercialization opportunity.

2. Results and Discussion

2.1. Structural Design Principle

Under bending strain, conventional stacked batteries (Figure 1a) usually suffer catastrophic damage, leading to performance degrades or even causes failure. Conventional stacked batteries are faced with significant stress concentrations at bending centers during mechanical deformation (Figure 1b). When subjected to cyclic bending‐recovery strain conditions (Figure 1c), these architectures exhibit progressive mechanical degradation: internal components form a brittle fracture network (mechanical failure mode). Meanwhile, the electroactive layers undergo yield‐induced delamination (electrochemical failure mode). This synergistic damage culminates in eventually triggers catastrophic performance decay and potential functional failure of the energy storage system (Figure 1d).

FIGURE 1.

Design and fabrication schematic of fern‐like FLIB. (a–d) Schematic of conventional stacked battery failure under mechanical stress. (e) The extension–contraction of a fern. (f) Abstract extraction of design elements. (g) Bending of fern‐like battery nodes. (h) Analogy between the ferns and the fern‐like battery. (i) Schematic illustration and (j) actual fabrication process for the fern‐like battery.

Two morphological stages in fern growth are illustrated in Figure 1e. During germination, the pinnae converge radially along the rachis in a helically coiled configuration. As it matures, the rachis progressively uncoils while pinnae deploy radially from the central axis. Upon full expansion, the rachis assumes linear geometry with pinnae arranged in fractal phyllotaxis. Nature‐inspired design, the unique extension‐contraction behavior of ferns, provides an elegant strategy to address stress concentration challenges in conventional stacked batteries under bending deformation. Conceptualizing the growth process of ferns into a concrete model and translating the biological structure into a novel fern‐like flexible battery structure, namely the node deformation strategy (Figure 1f, h). Specifically, the main stem is equivalent to a series of connected hinges. At the same time, the leaves are regarded as horizontally unfolded energy storage components, thereby forming two design elements of flexible batteries: the bending expansion of the connecting components and the fractal arrangement of the energy storage components. When subjected to bending stress, the concentrated stresses are redistributed by the controlled nodal deformation (Figure 1g and Figure S1). In other words, the mechanical energy is dissipated through discrete node units, thereby maintaining low stress states within energy‐storage blocks. Meanwhile, the bending radius is increased, thus alleviating the excessive strain at the top region due to the discrete node units. Consequently, such a fern‐like design is conducive to retaining superior electrochemical performance under repeated alternating bending stresses.

Guided by nature‐inspired design strategies, the schematic illustration (Figure 1i), and actual fabrication process (Figure 1j) of biomimetic flexible batteries are systematically presented. Initially, the cathodes and anodes were punched into the fern‐like geometries, ensuring that both electrodes retained intact and consistent biomimetic segmented node structures. Notably, the punched fern‐like cathode and anode were strictly aligned in a coaxially symmetric manner to guarantee the structural integrity and electrochemical stability of the battery under repeated mechanical deformation. Specifically, the geometric centers of the fern‐shaped energy storage blocks on the cathode and anode were completely coincident, and the central axis of the connecting nodes was parallel to the length direction of the battery. Importantly, the alignment design enabled the cathode and anode to maintain a well‐matched opposition during repeated bending deformation, which is not reduce irreversible capacity loss caused by the deteriorated electrode matching, but also mitigates local stress concentration induced by uneven stacking during dynamic bending cycles.

Following this alignment strategy, the electrodes were stacked in the sequence of a single‐side‐coated cathode, separator, double‐side‐coated anode, separator, and single‐side‐coated cathode. Meanwhile, the separator was designed with a marginal excess relative to the electrodes, which further eliminates the risk of internal short circuits caused by minor assembly deviations and long‐term mechanical deformation. Finally, the entire assembled electrode stack was encapsulated using an aluminum‐laminated film to obtain the integrated fern‐like flexible battery.

Considering the operability of the assembly process and performance of the battery, the fern‐like battery node unit is designed as a 10‐array to analyze its performance and anti‐failure mechanism (Figure S2). The packaged battery using aluminum‐laminated film exhibits a maximum thickness of 0.647 mm (Figure S3), satisfying requirements for seamless integration in wearables and microbots. Figure S4 demonstrates a scalable electrode via modular unit adjustment, enabling application‐specific customization. This bioinspired nodal design framework transcends specific geometries and accommodates diverse electrode dimensions/active materials.

2.2. Electrochemical Performance

FLIBs are typically operated under three distinct operational modes: (i) Static mode, that is, maintaining predefined deformation when powering devices as structural components. (ii) Dynamic mode, which means powering devices after undergoing cyclic mechanical loading. (iii) In situ dynamic mode, where simultaneously sustaining shape transformation and power delivery. These modes provide the foundation for a comprehensive electrochemical performance evaluation in all operational scenarios.

The static long‐term cycling performance is a key metric for evaluating the long‐term adaptability of FLIBs to strain. The fern‐inspired battery was maintained in the nominal bending (Figure 2a, R = 33 mm) and small‐radius bending (Figure 2b, R = 10 mm) during charge/discharge cycles at 0.5C to verify its electrochemical performance under static mode. To achieve consistent capacity, the initial stabilization cycles of FLIBs are performed under flat conditions, and then the batteries are continuously cycled under continuous nominal deformation. The battery can deliver a high capacity of 126.0 mAh g⁻¹ during the first cycle, and the capacity retention of approximately 93% is retained after 50 cycles, with nearly 99.7% CE (Coulombic efficiency) (Figure 2c). Remarkably, even under extreme mechanical constraint at a small‐bending radius, the capacity retention remains at 85.2% and high average CE (99.4%) after 50 cycles (Figure 2d), indicating that the FLIBs have excellent deformation tolerance and stable electrical contact. Since the bending‐induced degradation may remain undetected in conventional capacity tests, it often manifests prominently under high‐rate conditions. Rate capability under mechanical deformation was systematically evaluated in the nominal bending state (Figure 2e) and small‐radius bending state, respectively (Figure 2f). At nominal curvature, the battery delivered a high average discharge capacity of 139.3, 122.5, 117.6, and 111.5 mAh g⁻¹ at 0.1C, 0.25C, 0.5C, and 1C, respectively. When the rate is restored to 0.5C, the average discharge capacity can recover to 116.5 mAh g⁻¹. Under severe small‐radius bending constraints, the corresponding capacities are 125.6, 116.2, 110.8, and 101.0 mAh g⁻¹ at 0.1C, 0.25C, 0.5C, and 1C, respectively. The outstanding capacity of 107.0 mAh g⁻¹ can be achieved after a fresh 0.5C rate cycle, confirming exceptional electrochemical reversibility even under severe small‐radius bending and high rate. In addition, when bent at a nominal radius, the battery exhibited lower overpotential at different rates (Figure S5).

FIGURE 2.

Electrochemical performance of the fern‐like battery under static and dynamic modes. Digital images under (a) nominal and (b) small‐radius bending. Cycle stability under (c) nominal and (d) small‐radius bending. Rate capability under (e) nominal and (f) small‐radius bending. (g) Cycle stability of conventional stacked battery under sustained bending strain, (h) corresponding charge–discharge cycles, and (i) magnified view. (j) Cycle stability under nominal bending and small radius bending strain states in dynamic mode. (k) Voltage‐normalized capacity curve at the 35th, 60th, and 90th cycles during dynamic deformation, and (l) cycle stability of conventional stacked battery under nominal and small‐radius bending strains.

As a comparison, cycle stability experiments with conventional stacked batteries reveal intrinsic bending vulnerability (Figure 2g–i). The initial electrochemical performance remains stable for the first five cycles, but abnormal voltage oscillations occurred in the sixth cycle, characterized by abrupt N‐shaped voltage collapse/surge transitions (Figure 2i). In conventional stacked batteries, fracture of current collectors gives rise to electrical isolation of electrode segments and interfacial delamination, thus increasing contact resistance and inducing discontinuous voltage collapse during operation. Consequently, the capacity rapidly declines and eventually leads to a complete failure, which is consistent with the theoretical prediction.

The cyclic strain tolerance of the fern‐like battery was evaluated through dynamic mode testing (Figure 2j). Firstly, 10 charge–discharge cycles were performed under flat states to obtain cycle capacity. Following the flat state cycles, the battery underwent 2 000 bending cycles (from flat to nominal bending) on linear guides, and then 25 charge–discharge cycles under nominal bending. The discharge capacity in the first cycle was 117.9 mAh g⁻¹ and remained at 113.6 mAh g⁻¹ after 25 cycles. The decay of discharge capacity is only 0.017% per cycle, which indicates that the bending deformations have nearly no influence on the designed battery during this stage. The initial decrease in capacity during the bending cycle may be attributed to unstable interfacial contact within the battery. As the configuration adapts to the deformed state, electrode contact stabilizes, allowing capacity to recover. The battery was then returned to the flat configuration for 10 steady cycles. Subsequently, the battery underwent small‐radius bending over 500 times and remained in the small‐radius bending state. The battery was cycled for the next 15 cycles with stability, maintaining a capacity of 113.0 mAh g⁻¹. At the end of small‐radius bending, the battery was continuously charged and discharged 10 times in flat states. During this stage, the specific capacity decreased from 113.0 to 109.7 mAh g⁻¹. This capacity fade is likely attributable to strain release upon returning to a flat configuration. The resulting interfacial instability facilitates relative displacement between the cathode, separator, and anode, compromising contact integrity and leading to reduced capacity. As the last step, the battery was returned to the nominal bending configuration for 20 cycles, with the capacity declining only marginally to 109.0 mAh g⁻¹. In general, after withstanding over 4 500 mechanical bends and 90 charge–discharge cycles, the designed battery performs steadily and maintains about 92.4% of the initial capacity, with an average coulombic efficiency of better than 99.4%, indicating that almost no side reactions and no obvious mechanical degradation occurred during the tests.

The selected last curves from each dynamic stage (besides the flat state) are shown in Figure 2k. The voltage plateaus remain almost unchanged and only slight changes in the capacity profile, which suggests that the fern‐like battery had a steady energy output under dynamic conditions. Figure 2l presents the cycle performance of a conventional stacked battery subjected to identical mechanical bending. The stacked battery exhibits significantly poorer capacity stability than the fern‐like battery after only 20 cycles. Following subsequent small‐radius bending, the capacity of the stacked battery becomes increasingly unstable until eventual failure. As shown in Figure S6, the charge–discharge curves of the conventional stacked battery after 500 small‐radius bending cycles lose their smooth profile and develop a pronounced zigzag shape. This deterioration likely stems from structural damage to the electrodes induced by mechanical bending, where electrode burrs penetrate the separator, causing micro‐shorts. It should be noted that the FLIBs maintain robust operation even under small‐radius bending, benefiting from effectively managing localized strain in the fern‐like battery architecture.

Practical engineering applications of FLIBs demand not only stable electrochemical performance under static and dynamic configurations but also consistent power output during in situ dynamic operation. Therefore, the FLIBs are constantly bent and deformed during the charging and discharging cycles, and the cycling performance of the fern‐like battery under in situ dynamic bending is demonstrated in Figure 3. The FLIB is sequentially subjected to a multi‑stage dynamic bending process by the actuator (Figure 3a,b). Stage I involves deformation from the flat state to nominal bending, while Stage II corresponds to further bending to a small radius. The specific configuration of the testing system and battery dynamic movement videos is documented in Videos S1 and S2. Figure 3c depicts the capacity retention of the designed battery throughout approximately 42 000 bending cycles over 60 charge–discharge cycles. The initial cycle comprised 10 cycles in a flat configuration to stabilize capacity. Subsequently, the battery underwent Stage I, which involved 21 000 concurrent bending deformations over 20 cycles. Importantly, no capacity degradation is observed during this stage. Instead, a slight capacity increase appears, which is attributed to the improved electrolyte wetting and enhanced interfacial contact between electrode particles and the current collector induced by mechanical deformation. Following Stage I, the battery is returned to a flat configuration for 10 stabilization cycles to eliminate residual strain. Finally, the battery entered Stage II, performing 20 cycles with 21 000 small‐radius bending deformations. Surprisingly, upon completion of the in situ dynamic cycle protocol, the battery only exhibited a 4.3% reduction in capacity, corresponding to an average capacity loss of approximately 0.0023‰ per bending deformation. Such outstanding cycling performance indicates that, even under severe dynamic deformations, the unique battery design inspired by fern plant leaves can still effectively preserve the integrity of key battery components, including active materials, current collectors, and electrical contact.

FIGURE 3.

Electrochemical performance of fern‐like batteries under in situ dynamic bending. (a) In situ dynamic testing of batteries‐optical photos of the mechanical strain system and (b) schematic diagram of the bending testing process. (c) Cycle stability and coulombic efficiency. (d) EIS in the flat state and after dynamic bending cycles. (e) Galvanostatic charge–discharge profiles at cycles 1, 13, and 43 during dynamic deformation. (f) Comparative analysis of cycle times, mechanical durability (maximum deformation times), and capacity degradation between fern‐like batteries and reported FLIBs. Conventional stacked battery (g) flat discharge and (h) in situ dynamic mode charge/discharge curves with enlarged sections.

The electrochemical impedance spectroscopy (EIS) was employed to assess the interfacial stability and charge transfer kinetics of the fern‐like battery after in situ dynamic cycling (Figure 3d). After the cycling, the real‐axis intercept of the Nyquist plot showed no significant alteration, indicating that dynamic bending did not compromise the solution resistance or internal resistance of the electrodes. This demonstrates the inherent stability of the fern‐like design in maintaining electrical contact and flexibility. The diameter of the high‐frequency semicircle is primarily governed by the charge transfer resistance. A slight increase after each cycle suggests a minor increase in the electrode reaction resistance during charge/discharge, which is likely attributed to subtle interfacial ion transport effects induced by the mechanical deformation. Accordingly, after 42 000 dynamic bending cycles, no abrupt changes are observed in the EIS parameters, underscoring the long‐term cycle stability of the fern‐like battery under flexible operating conditions. As expected, the galvanostatic charge/discharge profiles (Figure 3e) remained stable and virtually unchanged in shape across flat (first), Stage I (13th), and Stage II (43rd) configurations, indicating excellent preservation of the battery's internal structure. Furthermore, the overall and locally magnified discharge voltage profiles (Figure S7) confirmed high electrochemical stability even under aggressive deformation. Furthermore, as shown in Figure 3f, the fern‐like battery demonstrates a competitive electrochemical stability and robust mechanical durability compared to other reported flexible lithium‐ion batteries [15, 21, 22, 23, 24, 25, 26, 27, 28]. The excellent in situ dynamic performance suggests that the battery system can effectively self‐adjust to adapt to external stresses during continuous bending deformation, and maintain high electrochemical performance even in tight bending states with higher stress concentration.

In contrast, conventional stacked batteries subjected to identical in situ bending strain cycles exhibited catastrophic failure (Figure 3g and Figure S8). Once the dynamic bending deformation is performed, the discharge voltage collapsed abruptly before cycle completion. The enlarged voltage profiles (Figure 3h) show that within the 11 cycles over a period of 40 min, there are severe and large‐amplitude oscillations, indicating structural degradation and mechanical failure. The repeated tension‐compression cycles during bending caused particle cracking/delamination in the cathode‐separator‐anode stack, disrupting local conductive networks and impeding electron transport, just as the instantaneous voltage increases (positive spikes). Under such conditions, the cathode‐anode contact caused micro‐shorts, leading to negative voltage spikes. Subsequent cyclic stress relaxation and recontact then give rise to high‐frequency voltage oscillations.

2.3. Mechanical Analysis and Deformation Mechanism

To further understand the mechanical properties of the bio‐inspired fern‐like battery design, finite element analysis (FEA) was performed to illustrate its electrode protection mechanism. The stress evolution in conventional stacked batteries and fern‐like batteries is compared in Figure 4a,b. Under a nominal bending deformation (radius R = 33 mm), the maximum principal stress of the conventional battery reaches 136 MPa, exceeding the yield strength of the aluminum foil (124 MPa) and causing irreversible plastic deformation. In sharp contrast, the stress in the fern‐like battery is effectively dispersed across the interconnected nodal regions, with the maximum stress drastically reduced to only 24.8 MPa. Even under extremely small‐radius bending (R = 10 mm), the maximum stress reaches only 52 MPa, safely below the current collector's yield strength (Figure 4c).

FIGURE 4.

Mechanical analysis and deformation mechanism of the fern‐like battery. Simulated stress distribution in a (a) conventional stacked battery and (b) the bioinspired fern‐like battery under nominal bending, and (c) the bioinspired fern‐like flexible battery under small‐radius bending. Schematic diagrams of deformation modes of the bioinspired fern‐like battery under (d) nominal bending and (e) small‐radius bending. (f) The node deformation schematics and (g) corresponding stress distributions. (h) Deformation process of the fern‐like battery under nominal bending and small‐radius bending.

To fundamentally elucidate the underlying stress dispersion mechanism, a segmented variable‐stiffness composite beam model was established without invoking small‐deflection approximations. The battery is theoretically abstracted as an alternating structure composed of high‐stiffness energy storage blocks (EI _Esb) and low‐stiffness flexible nodes (EI _Node). On the basis of the moment‐curvature constitutive relation, the curvature distribution follows $\frac{{\kappa }_{Node}(s)}{{\kappa }_{Esb}(s)}=\frac{E{I}_{Esb}}{E{I}_{Node}}\gg 1$. This mathematical relation dictates that, under a global bending moment, the bending curvature is predominantly concentrated within the soft node regions (κ _Node), whereas the high‐stiffness blocks undergo negligible deformation (k _Esb ≈ 0), thus intrinsically suppressing fracture of the active layers induced by normal stress. Since the high‐stiffness energy storage block experiences negligible bending, the entire bending angle θ is uniformly accommodated by n flexible nodes, with each node only undergoing only a small angular deflection of θ/n (Figure S9). Accordingly, the maximum stress (σ _Node,max) at each node is inversely proportional to n, that is ${\mathit{\sigma }}_{\mathit{Node},\mathit{max}}\propto \frac{\theta }{n·l_{Node}}$. In this manner, the fern‐like fractal design divides the extreme macroscopic bending curvature into sufficiently small and safe local strains across multiple micro‐hinges. By distributing the global bending angle θ among multiple segmented nodes, the large curvature that would otherwise localize at a single bending site is uniformly redistributed, which is the mechanical essence of the stress‐dispersion mechanism.

Optical images after dynamic bending, along with the disassembled electrode and separator (Figure S10) reveal that distinct wrinkle formation at the interconnecting nodal regions (as highlighted in the extracted node unit, Figure S1), confirming that the macroscopic flexibility is achieved purely through micro‐unit nodal deformation. Such a design effectively converts the traditional stress‐concentrated deformation mode into a distributed nodal strain. Furthermore, attaching polyimide films on both sides of the electrode further protects the metal foils from random severe bending.

Figure 4d captures the flexible battery undergoing two distinct bending regimes. One is nominal bending under gentle curvature, reflecting its overall bendability, and the other is small‐radius bending, highlighting localized deformation at the limit of its flexibility. Figure 4e illustrates a schematic diagram of the extension and coiling of a natural fern frond. The nominal bending state recapitulates the naturally unfurled morphology of a fern frond, representing large‐scale deformation within a natural form. Meanwhile, the pronounced local curvature induced by small‐radius bending parallels the tightly coiled structure of a fern fiddlehead, with the core coiling region highlighted by a black dashed line. To elucidate the underlying deformation mechanics, an animation visualizing stem morphology changes at key nodes during bending is provided in Figure 4f. Correspondingly, Figure 4g presents magnified finite element snapshots of critical nodal units within the deformed regions under both bending states. These simulations quantify the evolving mechanical response: with decreasing bending radius, significantly heightened strain and stress concentrations develop at these key structural nodes, which characterizes the stress distribution pattern near the structural flexibility limit.

Figure 4h depicts the physical processes of the battery under mechanical loading, including stress release and recovery (Stage I) and stress concentration at nodal deformation zones (Stage II). Red reference lines illustrate the stress distribution and highlight the functional role of the node units. In stage I, the initially fern‐like battery undergoes sequential phases: stress loading, release, full relaxation, and structural recovery. Under stress loading, the battery deforms uniformly due to the presence of nodal units, which effectively disperse localized stresses. During stage II, the reduced nodes indicate that a significant morphological change within the battery, as verified by the characteristics of concentrated deformation. Furthermore, the battery's bending behavior evolves through distinct stages (Figure S11). The deformation approximates a symmetric arch shape in stage I. With increasing stress (Stage II), the bending morphology evolves into a more complex and localized compact ring‐like form, while maintaining symmetry along the height axis. This transition from globally symmetric deformation to localized compact deformation underscores the dominant role of stress concentration in the bending mechanical response of the battery.

The electrode disassembled from the fern‐like battery remains intact. No detachment of active material or crack formation is observed in the connection region under the dominant strain, even after in situ bending deformation (Figure S12a). This further confirms that the fern‐like structure provides sufficient protection for the electrodes. Correspondingly, SEM images of electrodes confirm that the junction region effectively retained its electroactive components, even under in situ dynamic deformation. In contrast, the electrodes of the conventional laminated battery both exhibit significant brittle fracture upon mechanical straining. Obvious detachment is also observed on the graphite electrode (Figure S12b).

It should be noted that the flexibility of batteries has certain limitations and cannot be sustained indefinitely bent. Based on the original fern‐like battery, double‐layer electrode (4 plates, 1.275 mm) and multi‐layer electrode (6 plates, 1.866 mm) assemblies were engineered to evaluate the flexibility limits. The corresponding electrochemical performance was characterized under in situ dynamic bending conditions (Figure 5a). Although the original single‐stack and double‐layer stacked (0.647 mm) batteries can cycle stably, the multi‐layer stacked battery declines rapidly. After only four cycles, they could no longer maintain normal charge/discharge capability (Figure 5b). The voltage profile also displays the pronounced “N”‐shaped fluctuations, characterized by sharp drops and rises in voltage (Figure 5c), suggesting severe structural failure inside the battery.

FIGURE 5.

Flexible limit of the fern‐like battery. (a) Cycle stability with different stacking thicknesses under sustained bending strain. (b) The charge–discharge curves of multi‐layer stacked fern‐like batteries, and (c) magnified view. (d) Mechanical model of multi‐layer stacked battery: (i) a simple model of a multi‐layer board, (ii) the static bending force distribution, (iii) the dynamic failure force distribution). (e) Stress distribution of interlayer failure in multi‐layer stacked batteries. (f) Schematic diagram of the stress distribution ahead of the crack in the graphite electrode under bending testing. (g) Geometry and boundary conditions of the graphite electrode notched specimen (g‐i) and the fracture process (g‐ii).

Furthermore, the multi‐layer stack is modeled as a laminated composite beam (Figure 5d‐i). By establishing a large‐deformation constitutive relation for the multi‐layer composite beam, theoretical calculations demonstrate that the cross‐sectional shear stress exhibits a parabolic distribution. Specifically, the shear stress vanishes at the upper and lower surfaces, while attaining its maximum magnitude precisely at the neutral axis (z = 0). Therefore, the longitudinal fracture is predicted to initiate preferentially within the middle graphite layer. Notably, the soft‐package battery introduces a critical deviation: the vacuum‐sealed structure generates an internal negative pressure (Figure 5d‐ii). According to interfacial confinement derivations, although this static pressure ( p ₀  = 1atm) effectively suppresses transverse interlayer delamination, it forces the low‐modulus graphite coating to absorb the extreme shear deformation. Consequently, horizontal shear stresses are generated between multilayer structures to resist deformation (Figure 5d‐iii), leading to the shear force concentrating and peaking within the midplane region, ultimately triggering shear‐induced failure of the graphite.

Particularly, the distribution of stress peaks is clearly elucidated via the FEA (Figure 5e). The color gradient across the cross‐section provides a direct visual representation of the stress intensity, where the red regions denote the maximum horizontal shear stress. This high‐stress band is concentrated exactly at the central layers of the battery stack, which visually validates the parabolic stress distribution predicted by theoretical modeling and pinpoints the dominant initiation site for structural failure. To reveal the underlying physical mechanism of this flexibility limit, the interaction between stack thickness and interfacial shear stress must be considered. Although the segmented node design effectively releases the bending normal stress (σ), the abrupt geometric transition at the node‐storage block interface introduces a severe stiffness discontinuity. According to the laminated composite beam theory [29], the nominal shear stress at the neutral axis is proportional to the square of the total thickness (τ_nom ∝h ²). Therefore, the increase in thickness (h) leads to a significant nonlinear amplification of the nominal shear stress. Furthermore, the amplified stress is further intensified by the local stress concentration factor (K _t,τ ) at the geometric step of the node interface. For the multi‐layer (6 plates, 1.866 mm) battery, the superimposition of the thickness‐squared effect and geometric discontinuity causes the local cyclic shear stress to rapidly exceed the fatigue limit of the graphite layer, ultimately resulting in longitudinal shear fracture at the intersection of the neutral axis and the node interface. All in all, the node design maximizes planar flexibility via dispersing normal stress, yet is constrained by interfacial shear fatigue, which ultimately determines the upper limit of vertical stacking layers for the battery.

Under in situ dynamic bending conditions, the interfacial shear stress exhibits periodic variation with the cyclic bending‐recovery motion, further accelerating fatigue damage accumulation at the graphite layer interfaces. Even if the induced stress in a single bending cycle is lower than the static delamination strength, repeated loading can still initiate and propagate microcracks along the interface following the Paris Law of fatigue crack growth (Figure 5f). Specifically, the localized stress field exists ahead of the crack tip. The concentrated stress continuously acts on the uncracked region, thereby driving the continual propagation of electrode cracks during cyclic bending.

To further verify the crack initiation and propagation behavior, FEA was performed to simulate the fracture process of the graphite electrode, as shown in Figure 5g. Figure 5g‐i illustrates the geometric configuration and boundary conditions of the graphite electrode sheets, where a pre‐existing notch is defined to simulate the vulnerable junction between the node and the energy storage block. Based on Neuber's elasticity theory, the geometric discontinuity at this junction acts as a V‐shaped notch, creating a severe stress concentration factor (K_t  > 1), which is also supported by the global finite element analysis in Figure S2. The corresponding numerical simulations of fracture behavior are presented in Figure 5g‐ii. The sequence of FEA contour plots dynamically visualizes the crack propagation process over time. The warm colors (red and yellow) represent the localized high‐stress concentration zone. As the sequence progresses, this high‐stress region continuously moves forward, reflecting the progressive yielding of the material and the steady propagation of the main crack through the graphite layer. Accordingly, the crack in the graphite electrode initiates at the node‐storage block interface, and then propagates along this critical boundary, as is further corroborated in Figure S13.

The multi‐layer stacked cells were disassembled after failure during in situ dynamic bending tests. The graphite electrode located exactly at the middle layer (neutral axis), especially the boundary between the flexible node and the energy storage block, exhibits severe longitudinal shear fractures (Figure S13a). Furthermore, a magnified view of the deformed region is shown in Figure S13b, no large‐scale delamination or fragmentation damage of the graphite occurred in this region, while the node region of the electrode is damaged owing to high stress concentration under small‐radius bending (Figure S13c). In detail, the distinct folds were formed on the surface of the graphite (Figure S13d) under the influence of the stress ripple effect, which can be regarded as a precursor feature for subsequent cracking. Obviously, the graphite layer underwent cracking and spalling along the junction between the node and the energy storage block, exposing the underlying copper foil current collector. Notably, the cathode electrodes on both sides exhibited excellent retention of structural integrity even at the junction between the node and the energy storage block (Figure S13e, f). The combination of the high elastic modulus of the cathode electrode and its positioning away from the region of maximum shear strain ensures that the cathode maintains structural integrity during dynamic bending. Therefore, the conductive network in this specific middle layer was entirely disrupted owing to the damaged graphite, which directly corresponds to the sudden “N‐shaped” voltage drops and rapid capacity decay observed in the electrochemical tests.

Guided by the failure mechanisms (shear stress concentration at the neutral axis and geometric notch effect), structural optimization strategies for flexible multi‐layer stacked batteries are proposed: (i) The optimization of the node‐energy storage block interface via fillet and gradient slot width designs (Neuber's solution) to reduce stress concentration. (ii) The introduction of an ultrathin sliding buffer layer between sub‐stacks to release shear strain. (iii) The prioritization of neutral axis engineering by using ductile materials instead of brittle ones (e.g., graphite) to withstand peak shear stress.

2.4. Flexibility Analysis and Application

The schematic diagram illustrating the flexibility characteristics of the designed battery is shown in Figure 6a. As the number of node units increases, the battery exhibits enhanced bending flexibility. The stacked battery structure features high regularity and compactness, where electrodes and components are arranged in rigid layered configurations. However, these layers exhibit mutual constraints with limited degrees of freedom, requiring the overcoming of substantial interlayer resistance during deformation, which consequently results in high rigidity and restricted bendability. As a significant contrast, the single‐node battery has only one node unit and possesses certain structural freedom, enabling localized deformation and limited bending near the node region. Nevertheless, its overall bending capability remains constrained. For the multi‐node battery, each node acts as a pivotal point, enabling concerted deformation under external force. The unique structural design endows the battery with greater structural freedom, improved shape adaptability, and thus a remarkable enhancement in overall flexibility.

FIGURE 6.

Flexibility analysis and application of the fern‐like battery. (a) Schematic illustrating the flexibility mechanism. (b) Relationship between the number of nodes, maximum radial deviation δ _max, and mechanical bending ease E _bend. (c) Relationship between relative effective bending stiffness and groove spacing. (d) Relative effective bending stiffness versus bending angle θ. (e) Demonstration of safe bending operation. (f) Front, back, and top–down views of a flexible display powered by an integrated fern‐like battery, demonstrating its deformation modes: flat, positive bending, negative bending, and arbitrary curvature. (g) A worm robot powered by an embodied fern‐like battery.

To quantify the enhanced flexibility of the fern‐like battery with increasing numbers of nodes, a combined geometric and mechanical perspective is essential (Figure 6b). Notably, the flexibility of the battery is mainly reflected in its bending morphology: when subjected to bending, the battery does not simply fold or bend at sharp angles, but rather can more accurately approach an ideal circular shape. An increase in the number of nodes can effectively promote a smoother and more continuous bending deformation of the battery, thereby significantly reducing the deviation from the ideal circular arc and lowering the mechanical resistance generated during the bending process. To accurately quantify the degree of fit between the battery surface and the circular arc surface, the maximum radial deviation (δ _max) was adopted as the evaluation index, which is defined as the maximum distance difference between a point on the battery surface and the ideal arc in the bending state. As the number of nodes is increased from n = 1 to n = 10, δ _max plummets from 246.7 to 2.5 mm. Obviously, a reduction of only two orders of magnitude in the radial deviation results in a 100‐fold increase in the surface adaptation efficiency, indicating that the battery has evolved from a state that was almost not circular (n = 1) to an ideal circle (n = 10). Therefore, the reasonable increase in the number of nodes can effectively enhance the bending flexibility of fern‐like batteries and optimize their surface adaptation performance.

Mechanically, increasing the number of nodes can significantly reduce the overall bending stiffness of the battery, thereby enhancing its deformability and enabling the battery to achieve a larger curvature (i.e., a smaller radius of curvature) under the same applied forces or moments. This enhanced deformability is quantified as the bending degree (E _bend), which increases exponentially with the number of nodes. Consequently, batteries with a larger number of nodes exhibit significantly softer bending behavior and can achieve high curvatures even under minimal external forces. The node design endows the battery with inherent mechanical compliance, allowing it to form a compact, near‐ideal circular structure.

Figure 6c illustrates the dominant influence of slot ratio (defined as groove width‐to‐cell width) on the mechanical properties of the FLIB structure. With the increase of the slot ratio, the relative effective bending stiffness exhibits a significant exponential decay. Notably, even a modest ratio of approximately 0.1 still induces a significant reduction in stiffness, indicating that excellent flexibility can be endowed to the battery without sacrificing excessive active material area. However, the mechanical advantage necessitates a critical balance with electrochemical reliability, because excessively narrow groove spacing markedly elevates localized stress (σ) within the battery structure (supported by Figure S2), which may further induce structural failure.

While the bending radius (R) serves as a direct parameter quantifying the overall deformation amplitude (encompassing both nominal bending and the small‐radius bending encountered during electrochemical testing), the bending angle (θ) presented in Figure 6d provides a refined analysis of the FLIB's flexibility limit (Figures S9 and S14). A sharp decline in bending stiffness is observed during the initial increase in θ (from 0° to 10°), whereas the response plateaus significantly beyond θ > 40°. Crucially, the inherent geometric correlation between θ and R ensures parametric self‐consistency within the entire system. This relationship is further validated in Figure 6e, which illustrates a distinct two‐stage evolution of the critical bending radius (i.e., the threshold at which the von Mises stress in the collector foil reaches its yield strength) as θ decreases: a precipitous drop in R from 33 to 10 mm occurs as θ decreases from 143° to 83°, followed by a more gradual decline to 5 mm as θ further reduces to 37.5°. This behavior clearly delineates the safe operational zone (stresses below yield strength) from the fracture‐prone zone. Significantly, the actual battery parameters (length L = 106 mm, bending diameter range 10–33 mm) consistently reside within the identified safe zone, thereby ensuring the mechanical integrity and electrochemical stability of the FLIB device.

A series of optical images integrated the fern‐like FLIB with a flexible LED display screen is presented in Figure 6f. Benefiting from its exceptional flexibility and stable cycle performance, the fern‐like FLIBs design enables it to stably power the LED display screen under various bending, even small radius bending conditions. Therefore, this integration constitutes a self‐powered flexible electronic device capable of stable operation under bending conditions. Given the battery's adaptability to surfaces of arbitrary curvature (Figure S15), it exhibits great potential for integration into a wide range of wearable and flexible electronic products.

In terms of outstanding flexibility, the fabricated fern‐like flexible battery is further applied to power a worm‐inspired robot (Figure 6g). During locomotion, driven by a crank‐slider mechanism, the FLIB fulfills dual functions: it acts as a structural support component directly integrated into the robot's moving body and simultaneously supplies power to the robot (Video S3). Encouragingly, the integration of FLIBs into the worm‐inspired robots not only reduces the robot's overall weight but also significantly enhances its endurance. Therefore, this approach successfully overcomes the conventional limitation that batteries are confined to functioning solely as non‐deformable structural elements. Furthermore, there is potential for integrating miniature devices (e.g., gas sensors) onto the battery. Deploying such robots in gas‐leak zones could enable hazardous gas detection, thereby enhancing safety for rescue personnel. The fern‐like FLIB, combining inherent flexibility and robust electrochemical stability under deformation, provides a versatile platform. It can serve as a conformal power source across a wide spectrum of flexible and wearable electronic systems.

3. Conclusions

In summary, a novel flexible battery was designed inspired by the flexible deployment/retraction mechanism of fern plants. The stress concentration is effectively mitigated through the redistribution of localized stresses across interconnected nodal units, thereby resolving critical stress issues on the battery's top surface. In particular, the fabrication of a fern‐like flexible battery requires only electrode punching and can utilize the established lithium‐ion battery manufacturing infrastructure, thus exhibiting strong potential for scalable engineering applications. Furthermore, the FLIBs exhibit superior cycling stability, and the capacity is retained at approximately 85.2% after 50 cycles under small‐radius bending. Additionally, the capacity retention is still 92.4% after more than 4500 mechanical bending cycles combined with 90 charge/discharge cycles in dynamic mode. Excitingly, only a 4.3% capacity decay after 42 000 concurrent bending deformations within 60 charge/discharge cycles (in situ dynamic mode). The stress‐distribution mechanism enabled by nodal deformation is systematically elucidated by FEA. Moreover, the fern‐like flexible battery exhibits outstanding overall performance and can be adapted to various mechanical deformations. Looking forward, this bio‐inspired, manufacturing‐friendly design paradigm, validated by exceptional stability under harsh dynamic conditions, paves the way for the development of truly durable and commercially viable flexible batteries, which are crucial for next‐generation wearable electronics and foldable devices.

Author Contributions

L.W., L.L.Z. and S.F.L. conceived the idea and supervised the project. L.W., L.L.Z. and S.F.L. designed the experiments and wrote the manuscript. S.L.G. and L.D. checks and modifies illustrations, participates in manuscript revisions. Y.Y.M. and Y.Z. draw a schematic diagram and provide corresponding suggestions for the manuscript. X.L. checked the manuscript. Experimental results were collected and analyzed through contributions of all authors.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File 1: smll73937‐sup‐0001‐SuppMat.docx.

Supporting File 2: smll73937‐sup‐0002‐VideoS1.mp4.

Supporting File 3: smll73937‐sup‐0003‐VideoS2.mp4.

Supporting File 4: smll73937‐sup‐0004‐VideoS3.mp4.

Supporting File 5: smll73937‐sup‐0005‐VideoS4.mp4.

Data Availability Statement

The data of this study are available from the corresponding authors upon reasonable request.