Skip to main content
Nature Communications logoLink to Nature Communications
. 2026 Sep 30;17:11256. doi: 10.1038/s41467-026-78129-0

Stress homogenization enabled by a layered composite electrolyte for high-performance all-solid-state lithium metal batteries

Qunyao Yuan 1, Zhe-Tao Sun 2,3, Gang Wang 1, Ce Zhang 1, Junyu Hou 1, Changsheng Song 1, Wu Sun 1, Yuanfan Gu 1, Tianke Zhu 1, Zhaoyu Wen 1, Xingyu Lei 1, Yun Song 1, Yongji Gong 4,5, Dalin Sun 1, Shou-Hang Bo 2,6, Xinyong Tao 7,8,✉, Jie Zhao 1,✉
PMCID: PMC13720145  PMID: 42957701

Abstract

The growing demand for high-energy-density and high-safety batteries has spurred research interest in all-solid-state lithium metal batteries. As the core constituent, solid polymer electrolytes exhibit improved interfacial compatibility compared with their inorganic counterparts. However, their further applications are impeded by suboptimal ionic conductivity and stress-induced uncontrolled Li deposition. Here, guided by in-situ stress imaging and multiphysical field simulations, we develop a layered composite electrolyte that simultaneously provides continuous ion-conduction pathways and modulates stress distribution for homogeneous Li deposition. Consequently, the layered electrolyte enables stable Li plating/stripping in a symmetric coin cell up to 3000 h with current density of 0.1 mA/cm2 and cutoff capacity of 0.05 mAh/cm2 at 60 °C. Most importantly, the Li|Layered electrolyte|LiFePO4 coin cell maintains a capacity retention of 84.3% after 400 cycles with 0.2C (34 mA/g) even at 30 °C. The corresponding pouch cell (nominal capacity of 7.65 mAh) delivers an initial discharge capacity of 6.3 mAh at 0.2C (1.53 mA) and 60 °C under a stack pressure of 100 kPa, achieving a capacity retention of 85.7% after 100 cycles. Our study reveals a strategy for utilizing structured solid electrolytes to modulate stress evolution and Li deposition for high-performance all-solid-state lithium metal batteries.

Subject terms: Batteries, Batteries, Batteries, Energy


Mechanical stress limits the lifetime of solid-state lithium metal batteries. Here, authors reveal that the mechanical electrochemical coupling governs battery cycling and develop a layered solid electrolyte that enables uniform lithium deposition and durable solid-state batteries.

Introduction

The growing demand for next-generation high-energy-density batteries has significantly spurred research on Li metal batteries (LMBs), due to their high specific capacity (3860 mAh/g) and low reduction potential (−3.04 V)1–4. In contrast to liquid batteries, solid-state counterparts offer enhanced safety by mitigating risks associated with leakage, flammability, and instability issues5–7. Among them, polymer-based solid-state batteries are particularly noteworthy for their favorable interfacial contact and easy processability8. However, in conventional polymer-based solid-state batteries, solid polymer electrolytes (SPEs) exhibit low Young’s modulus and yield strength, rendering them unable to resist elastic deformation9,10. Unfortunately, the Li metal electrode undergoes continuous and uncontrolled volume expansion during cycling. The pronounced volume change further compresses the electrolyte, leading to stress accumulation at the electrolyte-Li metal interface, which disrupts the Li-ion electroplating dynamics during cycling, resulting in uneven Li deposition and ultimately the growth of dendrites11.

Current strategies for alleviating stress-related issues are primarily focused on Li metal electrode engineering, such as artificial solid electrolyte interphases for modulating Li deposition12,13 and the design of Li metal electrodes with three-dimensional microstructures14 or soft substrates15,16 for mitigating interfacial stress. Another crucial solution to interfacial stress is the adoption of solid composite electrolytes to mechanically suppress the volume expansion of Li metal17–20. However, even robust electrolytes struggle to withstand the damage caused by such volume expansion, especially in the presence of Li dendrites, which have Young’s moduli ranging from 2.7 to 21 GPa21. Therefore, Kalnaus et al. emphasized that the rational design of solid-state electrolytes to effectively modulate the evolution of stress and strain is the key to the development of high-performance all-solid-state lithium metal batteries (ASSLMBs)11.

The evolution trajectories of Li polymorph morphologies and their electrochemical plating/stripping dynamics involve complex couplings among ion transport, electrochemical reaction, and the interfacial stress22,23. According to classical nucleation theory, the nucleation of Li metal is initiated by local fluctuations in composition and energy at the electrolyte-Li metal interface22,24. Thus, the distributions of the ion concentration and electric field profoundly influence Li nucleation and subsequent growth11. Meanwhile, the stress field significantly impacts the Li plating/stripping behavior and morphology of deposited Li polymorph. The reaction rate of Li plating is subject to the elastic energy (included in the Gibbs free energy) which arises from the compression and deformation of the electrolyte caused by the volume expansion of the Li metal25. Although some strategies partially mitigate the stress-related issues, the roots of stress evolution and propagation, as well as methods for the timely regulation of Li ion migration and deposition, remain elusive. Consequently, elucidating the mechano-electrochemical coupling relationships and investigating the mechanical and electrochemical properties of solid electrolytes can ultimately address challenges stemming from stress-related issues in ASSLMBs.

Here, in situ stress monitoring techniques are employed to investigate the stress evolution and distribution in polyethylene oxide (PEO)-based symmetric cells. The obvious stress concentration is revealed via the finite element method (FEM) analysis in the PEO-based electrolyte. The influence of stress concentration on Li-ion migration and deposition behavior is further investigated by decoupling mechano-electrochemical relationships based on a galvanostatic phase field model (PFM). On this basis, we propose a layered composite electrolyte (layered SPE) design strategy, featuring uniform stress distribution. As a proof of concept, a layered SPE is synthesized by integrating amorphous two-dimensional (2D) SiO2 nanosheets (SiO2 NSs) with the PEO-LiTFSI system (PEO-LiTFSI is denoted as bare SPE). As a result, the Li ion diffusion flux and deposition can be homogenized due to the uniform stress field derived from the well-designed structure of the layered SPE. As a result, the Li|Layered SPE|Li symmetric cell demonstrates Li plating/stripping stability of 3000 h at 60 °C and 700 h at 30 °C. Additionally, amorphous 2D SiO2 NSs exhibit abundant oxygen vacancies (Ovs) as Lewis acid sites, as well as the synergistic layered interpenetrating network, which affords a continuous supply of ion anchor sites26,27. These result in the formation of a highly effective ion transfer percolation network with ionic conductivity of 3.3 × 10−4 S/cm at 30 °C and ion transference number of 0.77 at 60 °C. When paired with LiFePO4 positive electrode, the all-solid-state full cell exhibits favorable capacity retention of 84.3% after 400 cycles with 0.2C (34 mA/g) even at 30 °C, which is a notable improvement in the PEO-LiTFSI system28–30. The pouch cell further demonstrates its potential for industrial applications with 85.7% capacity retention after 100 cycles at 60 °C.

Results

Stress visualization and multiphysical field coupling

To illuminate the stress evolution and distribution during Li plating and stripping processes, the in situ 2D stress monitoring and 3D real-time visualization techniques are employed to track the stress variation during Li plating/stripping. Firstly, a 50 mm × 50 mm pouch cell is assembled with Li|Bare SPE|Li symmetric configuration (Fig. S1). Then, a 2D in situ stress monitoring technique is employed to monitor the stress changes within the battery. The detailed working principle of the test system is shown in Fig. S2. The system mainly consists of two substrates, a pressure controller and a stress sensor, as shown in Fig. S2a. Briefly, the stress monitoring system operates by placing the pouch cell between two fixed substrates under controlled pressure. As the pouch cell swells during Li plating/stripping, the resulting axial force is detected by a planar stress sensor and converted into electrical signals for real-time imaging and analysis (Fig. S2b). Herein, a force of 25 kg is applied to the pouch cell via the pressure control system (an average pressure of 100 kPa), which is consistent with the typical stacking pressure operated in most polymer-based solid-state batteries and therefore does not affect the electrochemical performance of the pouch cell during measurements31. The characterization is conducted in a constant-gap mode, where the axial expansion force generated by the volume expansion of Li metal is recorded by the stress sensor array. Accordingly, the expansion stress before and after Li plating/stripping on each sensor array can be obtained32. The pouch cell is pre-cycled for at least three cycles to ensure the test system reach a steady state32. Figure 1a shows the variations in the expansion force of the Li|Bare SPE|Li symmetric pouch cell during Li plating/stripping, which is statistically recorded. The results indicate that the battery undergoes continuous and rapid stress growth, indicating significant volume changes of Li metal during Li plating/stripping. After 20 cycles, the normalized axial expansion force detected by the stress sensor increases from 1.0 to 1.5, which may be attributed to the gradual thickening of the deposited Li metal. The corresponding 2D stress map can be used to quantify the stress distribution during Li plating/stripping. The pouch cell is divided into a 5 × 5 grid by the planar stress sensor to enable real-time monitoring of the stress evolution during Li plating/stripping. Because the constant-gap mode is employed, the axial expansion force generated by the volume of the Li metal is detected by the planar stress sensor, resulting in the generation of a 2D stress distribution map. As shown in Fig. 1b, compared with cycle 5, the stress of the pouch cell increases significantly after 20 cycles, indicating the obvious volume expansion of Li metal. In addition, local stress concentration is observed within the pouch cell, which is attributed to the nonuniform Li⁺ deposition during Li plating/stripping. For visual observation of stress growth and evolution, a 3D stress visualization system is implemented with a 50,000 s monitoring cycle (Figs. 1c and S3). After 50,000 s of Li plating/stripping, localized stress concentration is observed, which is consistent with the 2D stress monitoring results. These findings indicate that the symmetric pouch cell with bare SPE experiences significant stress fluctuations arising from uneven Li deposition. Such gradually accumulated stress will compromise the mechanical stability of the solid electrolyte, manifesting as deformation and strain gradients11.

Fig. 1. Stress monitoring and mechano-electrochemical coupling relationships.

Fig. 1

a In situ monitoring of expansion force variations in a Li|Bare solid polymer electrolyte (SPE)|Li symmetric pouch cell at 0.1 mA/cm2. b 2D stress distribution in the Li|Bare SPE|Li symmetric pouch cell at different cycles. c Stress monitoring in a Li|Bare SPE|Li symmetric pouch cell after 50,000 s Li plating/stripping via 3D stress visualization technique. d The finite element method (FEM) simulation result of the stress response property of bare SPE at different stress propagation periods (Time 1 indicates the initial stage. Time 2 indicates the middle stage). e Visualization of the stress response within a specific horizontal line (white dashed line in d) parallel to the X-axis based on FEM simulation in the bare SPE system. f Stress variation over time at the point where X = 0 on the white dashed line based on FEM simulation in the bare SPE system. g The schematic of the mechano-electrochemical coupling relationships in the bare SPE system. Phase field model (PFM) simulation results of galvanostatic cycling under the modulation of h bare SPE system and i layered SPE system. The simulation is conducted at an  current density of 0.2 mA/cm2 for 3600 s Li plating.

To clarify the impact of stress formed during Li plating/stripping on the electrolyte, the FEM is employed to study the stress response characteristics of the bare SPE in detail. Here, a 2D model of a bare SPE matrix is constructed, and a hemisphere is used to simulate the process of deposited Li polycrystal compressing the electrolyte. The detailed simulation information is shown in Fig. S4 and Tables S1, S2. As depicted in Fig. 1d, the compression caused by Li polycrystals leads to the localized deformation of the electrolyte due to the low Young’s modulus of bare SPE. Notably, significant stress concentration and stress distribution gradient have also been observed due to the hindered propagation of stress within the electrolyte. Given that the migration and plating of Li ions under interfacial stress is a dynamic process during Li plating/stripping, the stress evolution within the electrolyte is further quantified. Thus, for the visualization of stress evolution over time, a region along a horizontal line parallel to the X-axis (the white dashed line in Fig. 1d) within the polymer matrix is selected for further study. As depicted in Fig. 1e, this horizontal line serves as the new X-axis, with time and stress as the Y and Z axes, respectively. Thus, the whole trajectory from stress accumulation to attenuation along this line is elucidated via a 3D visualization. The 3D cloud plot reveals that stress remains confined within a narrow spatial range along the X-axis over time, further validating the stress concentration in bare SPE. Simultaneously, to quantify the full cycle of stress evolution, the point located at X = 0 on the white dashed line is also earmarked for further investigation. The result indicates that the stress undergoes an accumulation period lasting for a normalized time of 0.3, followed by an attenuation period lasting for a normalized time of 0.7 (Fig. 1f).

As previously mentioned, the transport and deposition behavior of Li ions occurs under the coupling of multiphysics fields33. The elastic strain or stress gradients of ceramic electrolytes induced by interface stress have been proven to influence the electrochemical deposition behavior of Li ions11,25. However, this issue has been largely overlooked in polymer electrolytes. The elastic energy density of polymer electrolytes is comparable to or even exceeds that of ceramic electrolytes, according to the corresponding comparison (Tables S3 and S4). A more in-depth exploration of the mechano-electrochemical coupling relationship during cycling is crucial for better understanding the electrochemical deposition behavior of Li ions. Therefore, to elucidate the mechanism of stress in regulating Li deposition, a galvanostatic PFM is developed to decouple the mechano-electrochemical relationships, building upon the works of Zhao and Bo34,35. In the PFM system, interfacial evolution is governed by the minimization of total free energy functional (F)35,36:

F=∫Ωfchemξ,ci+fgrad∇ξ+felsξ,ε+felecξ,φi,ci,σhdV 1

The total Gibbs free energy (F) includes four components: the chemical potential (fchem), the interfacial energy gradient (fgrad), the elastic energy contribution (fels), and the electrostatic contribution (felec). See the Supplementary Information for more details.

The evolution of the lithium metal phase is governed by a nonlinear Allen–Cahn equation incorporating Butler–Volmer kinetics:

∂ξ∂t=−Lσδfchem+δfgrad+δfelsδξ−Lηh′ξexp1−αFηRT−cLi+c0exp−αFηRT 2

Equation 2 consists of two parts: the first represents the thermodynamic contribution, while the second corresponds to the kinetic contribution. Generally, the galvanostatic constraint can be expressed as:

∫ΩnFVLi,m∂ξ∂tdA=Icell 3

where Icell is the overall current. The detailed information is provided in the supplementary information.

Therefore, the entire process of Li plating/stripping in all-solid-state batteries can be simulated based on the galvanostatic PFM. Firstly, a 2D model of a bare SPE is constructed for simulating the cycling process (Fig. S5). The model operates under a constant current density of 0.2 mA cm−2 for Li plating and stripping. The detailed simulation information is shown in Table S5. In the bare SPE system, the PFM results are shown in Fig. 1g, h. Figure 1g presents a schematic to clarify the simulation results and the mechano-electrochemical coupling relationships. Generally, during the initial cycling, as Li deposition proceeds, the thickness of the Li metal electrode continuously increases. On the Li metal electrode side, the deposition is generally uneven, with surface protrusions exerting additional compression on the electrolyte, which leads to localized stress accumulation and a stress gradient in the bare SPE (stress field). This localized stress gradient further induces the redistribution of ions within SPE (ion concentration field), thereby disrupting the flux of Li ions11. The Li electroplating process, which involves both ion transport and electron transfer, is significantly driven by the local stress field. The detailed contributions of multiphysics fields to the mechano-electrochemical coupling are presented in Table S6. These findings highlight the significant role of elastic energy associated with interfacial stress in the mechano-electrochemical coupling during Li plating/stripping. As a result, the local stress field disrupts the Li plating/stripping, leading to inhomogeneous Li deposition morphology and the resulting uneven electric field25,37. The simulated distributions of the stress field, Li ion concentration field, and electric field based on the PFM are shown in Fig. 1h. In addition, the detailed time-dependent evolution of Li plating/stripping in the bare SPE system is shown in Figs. S6 and S7.

To address the issue of nonuniform stress distribution in bare SPE, it is imperative to introduce stress propagation media and design their arrangement for rapid stress response and uniform distribution. Natural nacre, with inorganic-organic composite structure, serves as a valuable study model. This biomaterial features a layered structure with the stacking of rigid 2D inorganic NSs within the polymer matrix, exemplifying the combination of high strength and rapid stress propagation38. Its mechanical response characteristics are further investigated via FEM simulation. Firstly, a 2D layered SPE model is constructed by introducing staggered SiO₂ NSs into the polymer matrix, while a hemispherical model is employed to simulate the compression of the electrolyte by deposited Li polycrystals (Fig. S8). The simulation results reveal that when localized stress is applied to the layered structure model, the rigid 2D SiO₂ NSs act as stress propagation media, uniformly dispersing the stress throughout the polymer matrix (Fig. S9a). Similarly, all points along a white dashed line parallel to the X-axis in Fig. S9a are selected to study the stress propagation process along this line. As depicted in Fig. S9b, unlike the stress concentration observed in bare SPE, the stress distribution in the layered structure exhibits a more uniform distribution along the X-axis. Additionally, compared to bare SPE, the response time for stress formation and dissipation in the layered structure is significantly shorter (Fig. S9c). These results indicate that the 2D SiO2 NSs, coupled with their specific arrangement, act as a medium for rapid stress propagation, which enables stress to redistribute rapidly and uniformly throughout the whole polymer matrix. As illustrated in Fig. 1i, the galvanostatic PFM results reveal that surface protrusions on the Li metal induce compression in the electrolyte. However, the rapid stress propagation and enhanced stiffness of the layered electrolyte facilitate uniform stress distribution. Compared to bare SPE, the layered SPE exhibits more uniform Li-ion concentration distribution, due to the uniform stress distribution and smaller stress gradients25. This uniformity ensures a stable Li stripping and plating process, ultimately achieving a homogeneous Li deposition morphology, as shown in Figs. S10 and S11. The detailed contributions of multiphysics fields to the mechano-electrochemical coupling are presented in Table S7. Given that electrolytes with lower lithium-salt concentrations are frequently reported in the literature, we have included supplementary simulation results utilizing a 1.2 M Li-ion concentration. As depicted in Fig. S12 and Table S8, the simulations for the bare SPE indicate that a non-uniform stress field exacerbates inhomogeneous Li⁺ deposition. Conversely, under the lower-concentration condition, the layered SPE system maintains highly uniform Li⁺ deposition during both plating and stripping (Fig. S13 and Table S9). Ultimately, these findings confirm that the stress-homogenization strategy enabled by the layered SPE is highly robust and broadly applicable across a wide range of Li-ion concentrations.

Design and characterization of layered SPE

As shown in Fig. 2a, a micro-to-macro multiscale construction strategy is employed to fabricate the layered SPE. Due to the passive properties of SiO2, it is widely used as an inorganic filler in SPEs to prevent side reactions with highly reactive Li metal. In Figs. 2b and S14, we synthesize amorphous 2D SiO2 NSs on a large scale via a wet chemical method. The scanning electron microscope (SEM) image shows the morphology of 2D SiO2 NSs with around 5 μm diameters (Figs. 2c and S15). Atomic force microscopy (AFM) reveals their smooth surface and the thin thickness down to 3.8 nm (Fig. 2d). Transmission electron microscopy (TEM) indicates the flexibility of 2D SiO2 NSs, and the corresponding Si element mapping shows the successful synthesis of SiO2 NSs (Figs. 2e and S16). Additionally, the high-resolution TEM (HRTEM) confirms the amorphous characteristic with disordered atomic arrangement (Fig. 2f). The amorphous properties of 2D SiO2 NSs are further characterized with X-ray diffraction (XRD), Raman spectroscopy, and electron paramagnetic resonance (EPR), as detailed in Figs. S17–19. The XRD pattern shows the absence of SiO2 crystalline peaks, but a broad peak is found around 22°. Raman spectroscopy reveals two characteristic peaks at 606 and 495 cm−1, which are indicative of amorphous SiO239. As the indicator of free electrons and oxygen defect sites, the EPR signal at g = 2.003 further confirms the abundant Ovs anchored on the amorphous 2D SiO2 NSs40. In addition, the mechanical properties of SiO2 NSs are also characterized via a microscopic nanoindentation characterization, and the detailed test protocol is shown in Fig. S2041. The Young’s modulus of SiO2 NSs is determined to be 68.29 GPa by combining nanoindentation and fitting, and their other mechanical parameters are also obtained as shown in Fig. S21 and Table S10. The high Young’s modulus of SiO2 NSs indicates their strong resistance to mechanical deformation41.

Fig. 2. Fabrication and characterization of the layered SPE.

Fig. 2

a Schematic illustration of the layered SPE featuring amorphous tablets and multiscale crosslinked interfaces. b Optical photograph of large-scale synthesis of 2D amorphous SiO2 NSs. Scanning electron microscopy (SEM) (c), atomic force microscopy (AFM) (d), and transmission electron microscopy (TEM) (e) images of SiO2 NSs, showing thin features. f High-resolution TEM image of SiO2 NSs. The insert is the selected area electron diffraction (SAED) pattern of amorphous SiO2 NSs. g Digital photograph of a 1-m-long roll of the layered SPE film. h Cross-section SEM image of the layered SPE film, showing a densely interlaced, homogeneously stacked and layered arrangement. i Corresponding Si element energy-dispersive X-ray spectroscopy (EDX) mapping.

The layered SPE is prepared by integrating 2D SiO2 NSs with widely adopted PEO-LiTFSI polymer electrolyte system. The nanoscale inorganic nanosheets-polymer composite units form an ordered, dense, and layered film via evaporation-assisted self-assembly38. Figure 2g displays the as-prepared 1-m-long roll of layered SPE film, showing its processability and potential for large-scale production. The thickness of the layered SPE is around 100 μm and can be easily tailored to specific shapes and sizes for subsequent cell assembly (Fig. S22). Additionally, the layered SPE exhibits flame-retardant properties, demonstrating high safety characteristics (Fig. S23). SEM is employed to characterize the cross-section of the layered SPE, and the image indicates the formation of an ordered, dense, and layered microscale structure (Figs. 2h and S24). The energy-dispersive X-ray spectroscopy (EDS) for Si elements confirms the uniform arrangement of 2D SiO2 NSs within the polymer matrix (Fig. 2i).

Micromechanical and macromechanical properties of SPEs

The mechanical properties of as-prepared electrolytes are systematically studied from both microscale and macroscale perspectives. To evaluate the electrolyte’s resistance to compressive deformation by Li polycrystals, the nanoindentation test is conducted along the Z-axis on the free sample surfaces, as shown in Fig. 3a. Firstly, Young’s modulus and hardness of the electrolyte films are evaluated by static nanoindentation characterization, as illustrated in Fig. 3b, c. The Young’s modulus (Ey) and hardness (H) of both bare SPE and layered SPE decrease slightly or stabilize relatively throughout the entire indentation depth. The average compressive Ey and H of bare SPE are confirmed to be 1.2 and 0.07 GPa, respectively. In contrast, the Ey and H for the layered SPE are significantly higher at 3.7 and 0.17 GPa, demonstrating its greater ability to resist elastic deformation than bare SPE19,42. Additionally, across the entire contact depth, the layered SPE consistently displays uniform Young’s modulus and hardness distribution, indicating homogeneity in its mechanical properties at the microscale. To further investigate the uniformity of the electrolyte over a larger area, a 60 μm × 60 μm region was selected for nanoindentation testing. Within this region, 16 testing points arranged in a 4 × 4 grid were measured to obtain the values of Young’s modulus and hardness, which were subsequently used to construct the 2D maps of modulus and hardness (Fig. S25). It should be noted that the lateral size of the SiO2 NSs shows negligible influence on the 2D modulus distribution, because the nanoindentation measurement characterizes the compressive modulus of the electrolyte38. As illustrated in Fig. 3d, e, the typical contour maps of Ey and H of layered SPE demonstrate its relatively stable and uniform Young’s modulus and hardness at the microscale, which is relatively higher than that of bare SPE (Fig. 3f, g), consistent with the above results of static nanoindentation. Additionally, the results of contour maps of Ey and H of bare SPE are also consistent with static nanoindentation. The Young’s moduli of electrolytes are further investigated using AFM. According to the Young’s modulus mapping, layered SPE displays a mean value of 3.1 GPa, significantly surpassing the 0.9 GPa measured in bare SPE. This finding further confirms the enhanced mechanical performance of layered SPE (Fig. 3h, i).

Fig. 3. Micromechanical and macromechanical properties of layered SPE.

Fig. 3

a Representative scheme for one-direction (Z) probing of nanoindentation contact response. Young’s modulus (b) and hardness (c) of electrolyte films measured by static nanoindentation test. Typical contour maps of Young’s modulus distribution (d) and hardness distribution (e) for layered SPE obtained from a scanning area of 60 μm × 60 μm. Typical contour maps of Young’s modulus distribution (f) and hardness distribution (g) for bare SPE obtained from a scanning area of 60 μm × 60 μm. 3D AFM Young’s modulus mappings of layered SPE (h) and bare SPE (i). All the mechanical tests are conducted at 25 ± 2 °C with 25% relative humidity (RH).

The macroscopic mechanical properties of the electrolytes are also examined via uniaxial tensile tests. As shown in Fig. S26, the layered SPE demonstrates a relatively high tensile Young’s modulus of 190 MPa and a high yield strength of 20.5 MPa, which shows significant advantages compared to other reported works9,18,43. In contrast, the Young’s modulus and yield strength of the bare SPE are 80 and 4.1 MPa, respectively. It should be noted that there is a significant difference between the compressive modulus of the electrolytes obtained from nanoindentation and the tensile modulus obtained from uniaxial tensile. This difference arises from the inherent characteristics of glassy polymers and the microstructural anisotropy of the layered SPE44. In addition, the higher in-plane stiffness endows the layered SPE with greater capacity to resist shear deformation, thereby preventing Li-ion rearrangement induced by deformation and ensuring a uniform Li-ion distribution within the electrolyte11. Additionally, the layered SPE also exhibits extraordinary plastic deformation of up to 2250%, thus leading to high toughness of 483 MJ/m3, significantly higher than bare SPE with toughness of 70 MJ/m3. The plastic deformation and toughness of the electrolyte play a crucial role in enhancing cycling stability by delaying mechanical failure of the electrolyte and dissipating the stress accumulated at the electrolyte-electrode interface11,19,45. The toughness and yield strength of the layered SPE are also compared with previously reported mechanically enhanced polymer-based electrolytes (Fig. S27 and Table S11). The results show that the layered SPE not only possesses high yield strength but exhibits extraordinary toughness, significantly surpassing other similar polymer electrolytes.

Electrochemical characterization of layered SPE

The influence of adding amorphous 2D SiO2 NSs on the electrochemical performance of the layered SPE has been comprehensively investigated via the following various physicochemical characterizations. A series of electrochemical impedance spectroscopy (EIS) measurements are conducted based on the configuration of SS|Layered SPE-xSiO2|SS (SS and x denote the stainless steel and mass ratio of SiO2 to PEO, respectively). Figures 4a and S28 show the Arrhenius plots of layered electrolytes with different SiO2 contents based on the temperature-dependent ionic conductivities. With the addition of SiO2, the ionic conductivity of the electrolyte is significantly higher than that of bare SPE. However, when the mass ratio of SiO2 to PEO rises to 0.4, the ionic conductivity decreases due to excessive SiO2 content impeding the mobility of the polymer chains and thus hindering ion migration. In contrast, the optimized layered SPE-0.3SiO2 composite exhibits the highest ionic conductivity of 1.6 × 10−3 S/cm at 60 °C and 3.3 × 10−4 S/cm at 30 °C as well as the lowest activation energy of 0.39 eV, which is higher than that of bare SPE with a value of 2.3 × 10−4 S/cm at 60 °C and 2.7 × 10−5 S/cm at 30 °C as well as the activation energy of 0.53 eV (Fig. S29). Therefore, the mass ratio of SiO2 to PEO is set as 0.3. Unless otherwise specified, all composite electrolytes adopted in this work are prepared based on this ratio. The content of SiO2 NSs in the layered SPE-0.3 SiO2 is further confirmed by the thermogravimetric analysis (TGA) characterization due to the high thermal stability of SiO2. As shown in Fig. S30, the residue content in bare SPE and layered SPE-0.3 SiO2 remains stable at 5% and 23%, respectively, as the temperature increases to 600 °C46. Therefore, the content of SiO2 NSs in the layered SPE-0.3 SiO2 is confirmed to be 18%, which aligns with the intended filler content. Figure 4b presents the differential scanning calorimetry (DSC) curves for pure PEO, bare SPE (PEO-LiTFSI), and layered SPE (PEO-LiTFSI-0.3SiO2) across a temperature range from 20 to 80 °C. The endothermic peaks at 71 and 61 °C correspond to the melting temperatures (Tm) of PEO and bare SPE, respectively. For layered SPE, the Tm notably decreases to 44 °C, indicating its lower crystallinity, which is responsible for the increased ion conductivity46. The linear sweep voltammetry (LSV) test is performed to assess the electrochemical window of electrolytes (Fig. 4c). The results indicate that the layered SPE exhibits an oxidation potential of 4.7 V, which is higher than the 4.3 V for bare SPE.

Fig. 4. Electrochemical and physicochemical properties of layered SPE.

Fig. 4

a Arrhenius plots of the electrolytes. “Layered SPE-xSiO2” refers to the PEO-LiTFSI-xSiO2 system (Li+: EO = 1:16), and x refers to the mass ratio of SiO2 NSs to PEO. b Differential scanning calorimetry (DSC) profiles of electrolytes from 0 to 80 °C. c The linear sweep voltammetry (LSV) curves of Li|Layered SPE-0.3 SiO2|SS and Li|Bare SPE|SS cells at a scan rate of 1 mV/s at 60 °C. d Li+ transference number test and polarization curve of layered SPE-0.3 SiO2 with the inserted electrochemical impedance spectra (EIS) before and after the polarization tested at 60 °C. e 7Li solid-state nuclear magnetic resonance (NMR) spectra of SPEs and LiTFSI for comparison. f Raman spectra of electrolytes. g Fourier-transform infrared spectroscopy (FTIR) analysis of PEO, bare SPE, and layered SPE-0.3 SiO2, with enlarged views of the 1300–1400 and 1020–1180 cm−1 regions. h Density functional theory (DFT) simulations of the structures of the PEO chain and amorphous SiO2.

Figure 4d demonstrates an ion transference number as high as 0.77 for the layered SPE system, which is much higher than that of bare SPE (0.35) and many previously reported values (Fig. S31 and Table S12). Additionally, the EIS plot shows no significant differences before and after chronoamperometry (the inset of Fig. 4d). The 7Li solid-state nuclear magnetic resonance (NMR) experiment is also employed to explore the origin of the high ion transference number, as shown in Fig. 4e. Compared to bare SPE, the 7Li NMR signal of layered SPE shifted 0.11ppm to the low field, indicating a weakened PEO-Li+ interaction47–49. According to previous reports, there are two states for TFSI- anions in Raman spectroscopy, including uncoordinated TFSI- near 740 cm−1 denoted as the P1 band and ion clusters near 745 cm−1 denoted as the P2 band46,50. The area ratio of the P1 to P2 bands is used to assess the degree of dissociation between Li+ and TFSI−. As shown in Fig. 4f, the proportion of free TFSI− in layered SPE is significantly higher than in bare SPE, indicating a higher degree of TFSI− dissociation. Fourier transform infrared (FTIR) spectra of pure PEO, bare SPE, and layered SPE are further compared. As shown in Fig. 4g, compared to pure PEO, which exhibits a vibrational peak at 1360 cm−1, the peak is absent in layered SPE, indicating a reduction in the crystallinity of PEO47. In layered SPE, the peaks at 1059 and 1146 cm−1 shift to lower wave numbers, suggesting a weakened interaction between Li ions and the ether-oxygen groups47,51. To further elucidate the origins of the high ionic conductivity and ion transference number in the layered SPE system, density functional theory (DFT) calculations are employed to compare the binding energies between TFSI− and PEO (Supplementary Data 1–3), and between TFSI− and amorphous 2D SiO2 NSs (Supplementary Data 4 and 5). In the TFSI− and Ovs-containing amorphous SiO2 system, the oxygen atoms in TFSI− are strongly attracted to the silicon atoms on the SiO2 surface, yielding a higher adsorption energy of 3.11 eV. It suggests a preferential adsorption of TFSI− on the SiO2 surface, facilitating the release of TFSI− from Li-[TFSI]n46 (Fig. 4h). The charge differential density (CDD) calculation is also analyzed to reveal the impact of Ovs on charge distribution between TFSI- and the SiO2 surface (Supplementary Data 6 and 7). As exhibited in Fig. S32, an obvious electron transfer is observed in Ovs-containing SiO2 NSs, indicating the stronger adsorption between TFSI- and amorphous SiO2 NSs52,53. It should be noted that the DFT calculations are based on simplified models and may not fully capture the complexity of the practical system.

Overall, the enhanced electrochemical performance of the layered SPE can be partly attributed to the infiltration of amorphized PEO domains, facilitated by the addition of 2D SiO2 NSs. Furthermore, the abundant Lewis acid sites anchored on amorphous SiO2 NSs, which interact strongly with the TFSI- anions, liberating the positively charged Li ions and subsequently improving the ionic conductivity of the composite polymer electrolyte. Another pivotal factor is the interpenetrating structure formed by the 2D SiO2 NSs within the polymer that serves as a substantial charge reservoir, offering continuous anchor sites and ion-conducting pathways, ultimately facilitating an effective ion transfer percolation network across much longer distances.

Uniform Li ion deposition with multiphysical field regulation

To further validate the modulation of stress on Li ion transfer and deposition behaviors, a series of symmetric cells are assembled in the configurations as Li|Bare SPE|Li and Li|Layered SPE|Li. Their long-term Li plating/stripping performances are subsequently evaluated at both 60 and 30 °C. The results indicate that the Li|Layered SPE|Li symmetric cell enables stable Li plating and stripping for up to 3000 h with an overpotential as low as around 20 mV at the current density of 0.1 mA/cm2 at 60 °C (Fig. 5a). In contrast, the Li|Bare SPE|Li symmetric cell experiences a permanent short circuit after 300 h of Li plating and stripping with overpotential of 120 mV. The long-term electrochemical Li plating/stripping performance of the layered SPE is also evaluated at a higher current density of 0.2 mA/cm2, and the result indicates stable Li plating/stripping for 1200 h (Fig. S33a). Additionally, the Li|Layered SPE|Li symmetric cell is tested at 30 °C to evaluate the electrochemical durability of the layered SPE electrolyte. The result demonstrates its ability to stably plate and strip Li for at least 700 h, indicating its potential for practical application (Fig. 5b). Furthermore, to evaluate the long-term stability of the interface between layered SPE and the Li metal electrode, corresponding electrochemical impedance spectra are also conducted at various stages of calendar aging (Fig. S33b). The result indicates that the interfacial impedance remains relatively stable even after 30 days.

Fig. 5. Modulation of layered SPE in Li plating/stripping stability, morphology, and component evolution of Li metal electrode.

Fig. 5

Li plating/stripping stability of symmetric coin cells at 0.1 mA/cm2, cutoff capacity of 0.05 mA h/cm2 under 60 °C (a) and 30 °C (b). Ex situ SEM (c) and 3D AFM images (d) of the Li metal electrode after 300 h Li plating/stripping at 60 °C with 0.1 mA/cm2, obtained from the Li|Bare SPE|Li symmetric cell under Ar atmosphere. Ex situ SEM (e) and 3D AFM images (f) of the Li metal negative electrode after 300 h Li plating/stripping at 60 °C with 0.1 mA/cm2 obtained from the Li|Layered SPE|Li symmetric cell. g Li 1s and O 1s X-ray photoelectron spectroscopy (XPS) spectra were conducted on the Li metal electrode surface after 300 h Li plating/stripping. h In situ monitoring of internal expansion force variations in a 50 mm × 50 mm Li|Layered SPE|Li symmetric pouch cell at 30 °C with 0.1 mA/cm2. i Stress maps for a Li|Layered SPE|Li pouch cell at different cycles at 30 °C with 0.1 mA/cm2. j Visualization of stress growth in a Li|Layered SPE|Li pouch cell plating/stripping at 30 °C with 0.1 mA/cm2 via a 3D real-time visualization technique.

Subsequently, the comprehensive properties of Li-containing compounds deposited on the Li negative electrode are investigated after 300 cycles. According to ex situ SEM images, severe Li dendrite growth and cracks are found on the Li negative electrode, implying inhomogeneous Li deposition and undulating surface morphology in the bare SPE system (Figs. 5c and S34a). This Li negative electrode is further investigated via AFM under inert atmosphere. As shown in Fig. 5d, the height mapping result reveals obvious variations in Li deposition morphology under the regulation of bare SPE system, with numerous cracks of varying depths, which is consistent with the morphology observed in the SEM image. For the bare SPE system, the Young’s modulus and energy dissipation properties of the deposited Li are further characterized. The average Young’s modulus is found to be 3.60 GPa, with values ranging from 0 to 5.00 GPa, reflecting the significant fluctuations in morphology and uneven distribution of the deposits. During the loading process of the AFM probe, the mechanical work performed is divided into dissipated energy and elastic recovery energy during unloading. Consequently, the surface morphology of deposits, such as surface undulations and microcracks, engenders significant influence on energy dissipation54. The energy dissipation mapping result of bare SPE also reveals the uneven Li deposition, aligning with the aforementioned height and modulus mapping results. In contrast, under the regulation of layered SPE, the deposition morphology on the Li negative electrode exhibits dense and smooth characteristics, implying uniform Li deposition behavior according to the SEM image (Figs. 5e and S34b). The height mapping on the Li negative electrode shows that the surface roughness of the deposits is as low as 40 nm, further indicating its smooth and dense surface. Subsequently, Young’s modulus mapping exhibits negligible fluctuations, with an average value of 7.60 GPa. This discrepancy could be attributed to variations in the morphology and compositional characteristics of the surface deposits. The energy dissipation mapping further verifies the homogeneous Li deposition on the negative electrode (Fig. 5f).

To investigate the chemical composition of the deposits on the Li metal electrode, ex situ X-ray photoelectron spectroscopy (XPS) characterizations are conducted after Li plating/stripping in Li|electrolyte|Li symmetric cells to detect the chemical composition of the cycled Li metal surface. Figures 5g and S35 reveal that the C1s spectra from both the bare SPE and layered SPE systems consist of peaks corresponding to carbide at 283.0 eV, C-C/C-H at 284.8 eV, C-OR at 286.3 eV, and Li2CO3 at 290.9 eV55,56. Upon further analysis of the O1s spectrum, in the bare SPE system, Li2CO3 (531.0 eV) is the primary component, followed by LiOH (529.6 eV) and a small amount of Li2O (528.6 eV)57. However, in the SiO2-modified system, the content of Li2CO3 significantly decreases, and Li2O becomes the dominant component. The Li2CO3 (54.9 eV) and Li2O (54.0 eV) signals are also observed in the Li 1s spectra from both bare SPE and layered SPE. In general, the Li-containing deposits formed from the bare SPE system primarily consist of Li2CO3, which is regarded as the passive layer and finally leads to high overpotential during Li plating and stripping47,58,59. However, for the layered SPE system, the primary component of the deposits is Li2O, which may result from the modulation of abundant Ovs on amorphous 2D SiO2 NSs. The Ovs effectively adsorb TFSI- anions, facilitating the lithiation of oxygen in PEO and forming more ion-conductive Li2O46. As a result, the Li-ion migration across the negative electrode interface is enhanced, significantly reducing interface resistance and overpotential.

Figure 5h illustrates the variation of stress within the Li|Layered SPE|Li symmetric pouch cell during Li plating/stripping. After 20 cycles, a relatively slow stress growth is observed, with the normalized stress increasing slightly from 1.0 to 1.1, much lower than that of the bare SPE system. The 2D operando stress imaging similarly reveals a uniform stress increase in the Li|Layered SPE|Li, indicating a smaller volume expansion of the Li metal due to denser Li deposition (Fig. 5i). The 3D stress monitoring results also demonstrate a highly uniform stress distribution, accompanied by slower stress growth compared to the bare SPE system (Figs. 5j and S36). These findings suggest that, under the modulation of the layered SPE, the internal stress of the battery is smaller and more uniformly distributed, which may be related to the denser and more uniform Li deposition during Li plating/stripping.

Layered SPE in all-solid-state Li metal full cells

To assess the electrochemical performance of electrolytes, all-solid-state coin cells are assembled using the bare SPE and layered SPE with LiFePO4 (LFP) positive electrode and Li metal negative electrode, namely Li|Bare SPE|LFP and Li|Layered SPE|LFP. The rate performance of Li|Layered SPE|LFP is firstly tested from 0.2C (0.2C = 34 mA/g) to 2C (2C = 340 mA/g) at 60 °C. The capacity of the Li|Layered SPE|LFP coin cell gradually decreases as the current density increased, dropping from 169 mA h/g at 0.2C to 140 mA h/g at 1C (1C = 170 mA/g). Obviously, even at the high rate of 2C, it can still provide up to 115 mAh/g discharge capacity. Besides, when the current density returns to 0.2C, the discharge specific capacity is restored to 166 mA h/g, realizing a capacity recovery of 98.2%, reflecting its stable rate performance (Figs. 6a and S37). The good rate performance of Li|Layered SPE|LFP may be attributed to the introduction of SiO2 NSs, which significantly enhanced the ionic conductivity of the composite electrolyte.

Fig. 6. Electrochemical performance of all-solid-state Li metal full cells.

Fig. 6

a Rate performance of Li|Layered SPE|LiFePO₄ (LFP) and Li|Bare SPE|LFP coin cells at 60 °C (1C = 170 mA/g). b Cycling performance of Li|Bare SPE|LFP and Li|Layered SPE|LFP full cells at 1C under 60 °C. c Corresponding charge and discharge curves of the Li|Layered SPE|LFP coin cell at different cycle numbers at 60 °C. d Cycling performance of Li|Layered SPE|LFP and Li|Bare SPE|LFP at 0.2C (0.2C = 34 mA/g) under 30 °C. e Cycling performance of Li|Layered SPE|LFP pouch cell at 0.2C (1.53 mA) under 60 °C.

Notably, the discharge specific capacities of Li|Layered SPE|LFP at high rates (1C and 2C) exhibit a significant advantage compared with the Li|Bare SPE|LFP. The cycling performance of the all-solid-state batteries is further examined at a rate of 1C under 60 °C. The results reveal that the layered SPE endows the battery with favorable cycling stability and the 82.4% capacity retention after 500 cycles (Fig. 6b, c). The stable cycling performance indicates long-term stability of the Li metal negative electrode during repeated stripping and plating, both chemically and mechanically, which is attributed to the layered structure of the electrolyte that promotes uniform Li deposition. In contrast, the Li|Bare SPE|LFP battery exhibits rapid capacity degradation from 112 mAh/g to 60 mAh/g, retaining only 53.6% of its initial capacity after 150 cycles. The poor cycling performance and overcharge behavior of the Li|Bare SPE|LFP battery may be attributed to the inferior stability of the Li metal negative electrode. The cycling performance of Li|Layered SPE|LFP is further evaluated at 30 °C, revealing its stable cycling for up to 400 cycles at 0.2C, with a discharge capacity retention of 84.3%. The favorable electrochemical performance of the Li|Layered SPE|LFP coin cell at 30 °C highlights its potential in practical applications (Figs. 6d and S38). To further authenticate the practical and scalable feasibility of layered SPE, a 50 mm × 50 mm Li|Layered SPE|LFP pouch cell is also assembled, followed by the assessment of its cycling performance at 0.2C under 60 °C (Fig. S39). The result indicates that the pouch cell exhibits an initial discharge capacity of 6.3 mAh at 0.2C and 5.4 mAh at the 100th cycle, indicating its potential in future industrial applications (Fig. 6e). To evaluate the comprehensive electrochemical performance, the ion transference number, ionic conductivity, and full-cell cycle number of layered SPE at 30 °C are compared to that of state-of-the-art SPEs in previous work, as a Ragone plot in Fig. S40 and Table S12. In essence, these favorable electrochemical properties underscore the pivotal role of well-designed layered SPE in modulating the holistic battery performance, while also demonstrating its potential for scalable applications.

Discussion

In summary, the origin and evolution of stress in SPEs are systematically investigated via combining 2D and 3D in situ stress sensing and monitoring techniques with FEM simulations. The galvanostatic PFM is further developed to elucidate the mechanical-electrochemical coupling relationships during cycling. On this basis, we develop a layered SPE with well-designed structure and enhanced mechanical strength, which provides continuous ion transfer pathway for high ionic conductivity and modulates stress distribution for homogeneous Li deposition. The layered SPE exhibits high ionic conductivity of 3.3 × 10−4 S/cm even at 30 °C, as well as stable Li plating/stripping up to 3000 h. The Li|Layered SPE|LFP all-solid-state full cell demonstrates high capacity retention with 84.3% after 400 cycles at 30 °C, which is challenging to achieve with traditional polymer-based electrolytes. The pouch cell further validates its potential for industrial applications with favorable cycling performance. We believe that the stress sensing and monitoring employed in this study offers a promising perspective for stress visualization study. The tailored electrolyte structure design and the elucidation of mechanical-electrochemical coupling relationships serve as a guide for addressing the issues of stress accumulation and mechanical failure resulting from uneven Li deposition.

Methods

Materials preparation

Preparation of 2D SiO2 nanosheets

Amorphous 2D SiO2 NSs were synthesized via a sol-gel process under ambient conditions at 25 ± 2°C. Typically, 0.7 mL NaOH (≥99%, Shanghai Aladdin Bio-Chem Technology Co., Ltd.) aqueous solution (2 M) was firstly added to 50 mL of water. The mixture was kept stirring until thoroughly mixed. Then, after the addition of 0.5 mL of tetraethyloxysilane (TEOS) (≥99.5%, Shanghai Aladdin Bio-Chem Technology Co., Ltd.), 0.5 mL of ethyl-acetate (≥99%, Shanghai Aladdin Bio-Chem Technology Co., Ltd.) was introduced while stirring. The resulting blend was then stirred for 2.5 h at 25 ± 2 °C. Subsequently, the precipitate was obtained through centrifugation and further washed with water and acetonitrile (≥99.5%, Shanghai Aladdin Bio-Chem Technology Co., Ltd.) for at least three times. Afterward, the obtained solid sediment was dispersed in acetonitrile for further use. Graphene (Gaoxitech, with a monolayer graphene content of 99.5% and a concentration of 10 mg g−1) is used as the template for the preparation of SiO2 NS powder. The prepared SiO2 NS is mixed with graphene, and then the graphene is removed in an air atmosphere at 700 °C.

Fabrication of PEO-LiTFSI-SiO2 solid electrolytes (layered SPE)

The solid electrolytes were prepared via a solution casting method. Firstly, 440 mg PEO (99%, Mw = 600,000, Sigma-Aldrich) and 179 mg LiTFSI (99%, Shanghai Aladdin Bio-Chem Technology Co., Ltd.) (with a mole ratio of EO/Li = 16:1) were mixed in 4 mL acetonitrile (99.9%, Shanghai Aladdin Bio-Chem Technology Co., Ltd.), and a homogeneous solution was obtained via Thinky mixer (Thinky Corporation). Subsequently, the amorphous 2D SiO2 nanosheets dissolved in acetonitrile with different weight ratio (20%, 30%, and 40% to PEO) was also added into the PEO-based solution. The white solution was achieved after the mixture was thoroughly stirred. The electrolytes were firstly coated on release paper in the dry room (25 ± 2 °C, 25%RH) using the Doctor Blade and then placed in a vacuum oven at 60  °C for 24 h to remove most of the organic solvents and moisture. LiTFSI readily draws moisture from the air. Therefore, after vacuum drying, the electrolyte films were quickly transferred to an Ar-filled glove box, where they were further baked at 60 °C for at least 3 days with the assistance of the glove box’s gas circulation system (organic gas filtration and activated carbon) to remove any residual organic solvent and adsorbed moisture as much as possible. Then the electrolyte films were continuously stored in the Ar atmosphere glove box for subsequent cell assembly. PEO-LiTFSI electrolyte was also obtained through the same method with PEO-LiTFSI-SiO2 electrolyte film.

Fabrication of LiFePO4 positive electrode

LiFePO4 (LFP) powder (98.5%) and Super P (98%) were purchased from Canrd New Energy Technology Co., Ltd. Typically, 320 mg LFP powder and 40 mg Super P were continuously mixed in mortar for 1 h. Then the mixed powder was dispersed in 700 mg N-methyl pyrrolidinone (NMP) (99.5%, Sigma-Aldrich) followed by the addition of 40 mg polyvinylidene fluoride (PVDF) (98%, Mw = 600,000, Canrd New Energy Technology Co., Ltd.) to form a homogenous slurry using Thinky mixer (Thinky Corporation). After that, the slurry was bladed on carbon-coated Aluminum foil (Dodochem, 15 μm Al foil with 1 μm carbon coating layer) to render uniform coating, which was further dried in vacuum oven at 80 °C for at least 24 h. After being transferred to an argon-filled glove box, the positive electrode foils were cut into circles with diameter of 10 mm using an electrode disc punch (Hefei Kejing Material Technology Co., Ltd.). The loading of LFP for coin cells was designed to be around 1–1.5 mg cm−2.

Materials for fabricating pouch cell

The Li foil (≥99.5%) with a thickness of 200 μm was purchased from China Energy Lithium Co., Ltd. (Tianjin). The LFP positive electrode (single side coated, similar to the preparation method described above for the coin cell, with loading ~1.8 mg cm−2). The prepared LFP positive electrode, layered SPE, and bare SPE were cut into the sheets with the size of 50 × 50 mm2 using a precision cutter (Deli Corporation A4 model). The Li foil was cut to a size of 52 × 52 mm2 using a precision cutter. The layered SPE and bare SPE film were cut to 54 × 54 mm2 using a precision cutter to fully separate the negative electrode and positive electrode. After all electrodes and electrolytes were fabricated in a dry room, they were transferred to an argon-filled glove box for size cutting and cell assembly. Pouch cells were assembled in an argon-filled glovebox with H₂O and O₂ concentrations below 0.1 ppm. The positive electrode, electrolyte membrane, and Li metal negative electrode were sequentially stacked in an aligned configuration to form the electrode/electrolyte assembly. The stacked components were then transferred into aluminum-laminated pouch bags, and the edges of the pouch bags were heat-sealed (YLJ-HP300, Hefei Kejing Material Technology Co., Ltd.). For the fabrication process, no electrolyte injection or gas-release process was involved. During cycling, a constant external pressure of 100 kPa was applied using a mechanical fixture.

Electrochemical measurements

Ionic conductivity measurement

The ionic conductivity measurement was conducted on a Donghua electrochemical workstation (Donghua Analytical Instruments Co., Ltd.). EIS measurements were performed on two-electrode SS|electrolyte film|SS cells (SS = stainless steel) with an AC perturbation amplitude of 10 mV over a frequency range of 10 MHz to 0.01 Hz, using 10 data points per decade, under an open-circuit voltage in the potentiostatic mode. All cells were tested in a temperature-controlled chamber.

Electrochemical window

The electrochemical stability of the solid electrolyte was studied by linear sweep voltammetry (LSV) on Donghua electrochemical workstation (Donghua Analytical Instruments Co., Ltd.) at 60 °C in a temperature-controlled chamber, with a potential range from 3 to 5.5 V (vs. Li+/Li) at a scan-rate of 1 mV/s. The cutoff voltage for the electrochemical stability was determined by the voltage value corresponding to the intersection of a baseline parallel to the potential axis and the slope of the LSV curve.

Ion transference number

The Li+ transference number (t+) of the solid electrolyte was measured by a chronoamperometry test on a Li|PEO-LiTFSI-SiO2|Li symmetric cell with perturbation signal of 10 mV at 60 °C in a temperature-controlled chamber.

Symmetric cells

The interfacial characteristics of the polymer electrolyte and Li negative electrode were detected on Li|PEO-LiTFSI-SiO2|Li symmetric cells using a Neware battery system (Neware-technology Co., Ltd.) at a constant current density of 0.1 and 0.2 mA cm−2 for 0.5 h charging and 0.5 h discharging at 30 and 60 °C in a temperature-controlled chamber.

Cycling performance

The cycling performance of Li|electrolyte film|LFP coin cells based on layered electrolyte and bare electrolyte were tested on the Neware battery system (Neware-technology Co., Ltd.). The coin cells were tested at 30 and 60 °C in a temperature-controlled chamber with different specific currents in a potential range from 2.5 to 3.8 V (vs. Li+/Li).

Materials characterization

For the scanning electron microscopy (SEM) characterization: the morphology of SiO2 NSs, Li metal after stripping/plating, and the cross-sectional morphology of electrolytes were investigated with SEM (ZEISS Gemini SEM 300 with EDS mapping, operated at 10 kV). It should be noted that the coin cells were carefully disassembled in an Ar-filled glove box, and any residual polymer on the surface of the Li metal was carefully removed. The treated Li metal was then stored in an Ar-filled sealed container and rapidly transferred for SEM characterization to prevent surface oxidation. For the transmission electron microscopy (TEM): TEM (Thermo Fisher Talos F200S G2, operated at 200 kV) was employed to characterize the morphology and microstructure of SiO2 NSs. For the preparation of the TEM characterization, a small amount of the SiO2 NSs was dispersed in the ethanol solvent and sonicated for approximately 5 min to break up any agglomerates. Next, a few drops of the dispersion were placed onto a copper grid coated with a thin carbon film. The excess liquid was carefully removed by blotting with filter paper to leave a thin, uniform layer of the material. The sample was then dried in a vacuum chamber to remove any remaining solvent. For the atomic force microscopy (AFM): the morphology and the thickness of SiO2 NSs were characterized via AFM (Bruker, Dimension FastScan Bio). For the sample preparation, a dilute ethanol solution of SiO2 NSs was slowly drop-cast onto a clean silicon wafer, and the excess solution was removed. The sample was then dried in a vacuum chamber to remove any remaining solvent. For the AFM characterization of Li metal after plating/stripping, the preparation of the Li metal sample was the same as its SEM characterization. The entire characterization process was carried out under an Ar atmosphere to prevent oxidation of the Li metal. For the X-ray diffraction (XRD): XRD was conducted on a Bruker diffractometer using Cu Kα radiation with a 2θ range of 10–90° at a scanning rate of 5°/min. For the sample preparation, the SiO2 NSs powder was ground into fine and uniform powder. Then the powder was evenly placed onto the sample holder and gently pressed to form a smooth surface. For the X-ray photoelectron spectroscopy (XPS): XPS study of Li metal after stripping/plating was performed on an ESCALAB 250 Xi system with monochromatic 150 W Al Kα radiation. The cells were disassembled in an Ar-filled glovebox, and any residual polymer on the surface of the Li metal was carefully removed. The treated Li metal was then stored in an Ar-filled sealed container and rapidly transferred for XPS characterization. For the electron paramagnetic resonance (EPR): the EPR was conducted on Bruker EMX Plus for the characterization of oxygen vacancies on SiO2 NSs. For the differential scanning calorimetry (DSC): the DSC was operated on DSC300, NETZSCH from 0 to 100 °C with a heating rate of 10 °C/min in a nitrogen atmosphere. For the sample preparation, the prepared electrolyte films were cut into small pieces and stored in Ar-filled sealed container and rapidly transferred for DSC characterization. For the thermogravimetric analysis (TGA): TGA was conducted on Mettler Toledo from 100 to 600 °C with a heating rate of 10 °C/min in a nitrogen atmosphere. Before the test, the sample was held at 50 °C in an inert atmosphere chamber for 1 h to remove any moisture that may have been absorbed during the sample transfer process. For the 7Li solid-state nuclear magnetic resonance (NMR) spectra: 7Li solid-state NMR was conducted on Avance Neo 400WB, Bruker. For the sample preparation, the as-prepared electrolyte films were cut into small pieces and ground into powder. The powder was then stored in an Ar-filled sealed container and rapidly transferred for further characterization. Meanwhile, as a control, LiTFSI powder was used as a reference and tested in the same manner. For the Raman spectroscopy: Raman spectra of electrolytes were obtained on Horiba LabRAM in a frequency range from 100 to 3400 cm−1. For the Fourier transform infrared (FT-IR) spectra: FTIR experiments were performed on Thermo Scientific Nicolet iS20 in an attenuated total reflectance (ATR) mode.

Mechanical tests

Macro-mechanical performance test

Uniaxial tensile tests were conducted on a universal testing machine. The 100 μm-thick electrolyte films were cut into 2 mm × 10 mm pieces. The initial gauge length for all samples was fixed at 3 mm, and the tensile speed was 1 mm/min. All tests were performed at 25 ± 2 °C and RH 25%.

Micro-mechanical performance test

The micro-mechanical performance of the electrolytes was evaluated via static nanoindentation testing. Prior to nanoindentation characterization, electrolyte films with a thickness of 100 μm were cut into 10 mm × 10 mm pieces and then securely mounted onto glass slides for subsequent characterization analysis. The nanoindentation tests were conducted at 25 ± 2 °C and 25% RH, using a TI950 triboindenter (Hysitron, Minneapolis, MN, USA) fitted with a Berkovich diamond tip (radius of curvature approximately 100 nm). To prevent substrate and surface influences, the Young’s modulus and hardness of the samples were measured at a depth of about 10% of the total thickness. These properties were determined from the unloading curve of each sample utilizing the Oliver-Pharr method. For the measurement of 2D Young’s modulus and hardness distribution, 16 testing points were selected within a randomly chosen 60 μm × 60 μm region (arranged in a 4 × 4 grid), and nanoindentation tests were performed. Other testing conditions were consistent with the aforementioned nanoindentation tests.

Stress distribution test

The stress distribution of all symmetric pouch cells was tested using high-precision pressure and displacement sensors, combined with an adaptive pressure regulation control system based on multiple swelling measurement modes (SWE2110, Initial Energy Science & Technology (IEST) Co., Ltd.). The total stress and stress distribution of the pouch cell during the charging and discharging process were collected by a pressure film system (BPD1000, Initial Energy Science & Technology (IEST) Co., Ltd.) with a force sensing matrix. Both the Li| bare SPE|Li and the Li|layered SPE|Li symmetric pouch cells operated with 0.5 h charging and 0.5 h discharging per cycle at the current density of 0.1 mA/cm2 under 30 °C. The detailed characterization information is presented in the Supplementary Information.

Simulation method

The detailed theoretical simulations (mechanical FEM simulation, multiphysical field simulation, and DFT simulation) are presented in the Supplementary Information, including the theoretical foundations, model construction, parameter settings, and procedure settings.

Supplementary information

41467_2026_78129_MOESM2_ESM.pdf (181.6KB, pdf)

Description of Additional Supplementary Files

Supplementary Data 1–7 (42.9KB, zip)
Peer Review File (7.3MB, pdf)

Source data

Source Data (28.4MB, xlsx)

Acknowledgements

The authors thank all members of the research team for their support and valuable discussions. The authors also acknowledge Initial Energy Science & Technology (IEST) Co., Ltd. for their technical support in the stress distribution tests.

Author contributions

J.Z., X.T. and Q.Y. conceived the concept. Q.Y., G.W., C.Z. and J.H. carried out the materials synthesis and performed materials characterization and electrochemical measurements. Z.S. and S.B. conducted the galvanostatic phase-field simulations. C.Z., C.S., W.S., Y.-F.G., T.Z., Z.W. and X.L. assisted in the synthesis and characterization of electrode materials and data analysis. Y.-F.G. conducted TGA and nanoindentation tests. Y.G. assisted in the assembly of the pouch cells. Y.S., Y-F.G. and D.S. provided important experimental insights and suggestions on the electrochemical mechanism and the structure of the manuscript. J.Z., X.T. and Q.Y. co-wrote the paper. J.Z., X.T., D.S., Q.Y., Y.-F.G., C.Z., G.W., J.H., Z.S. and S.B. contributed to the proofreading of the manuscript. All authors discussed the results and commented on the manuscript.

Peer review

Peer review information

Nature Communications thanks the anonymous, reviewers for their contribution to the peer review of this work. A peer review file is available.

Funding

This study is financially supported by National Key R&D Program of China (2022YFB2502000), National Natural Science Foundation of China (52225208, 52471223, and 22561142235), Science and Technology Commission of Shanghai Municipality (23160714000), Shanghai International Science and Technology Partnership Project (23520750400), and China Postdoctoral Science Foundation (2022TQ0065 and 2023M730614).

Data availability

The data that support the findings and conclusions generated in this study are provided in the main article, the Supplementary Information, and the Supplementary Data. Source data are provided with this paper.

Competing interests

The authors declare no competing interests.

Footnotes

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Contributor Information

Xinyong Tao, Email: tao@zjut.edu.cn.

Jie Zhao, Email: jiezhao@fudan.edu.cn.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41467-026-78129-0.

References

  • 1.Lee, M. J. et al. Elastomeric electrolytes for high-energy solid-state lithium batteries. Nature601, 217–222 (2022). [DOI] [PubMed] [Google Scholar]
  • 2.Yang, C. et al. Copper-coordinated cellulose ion conductors for solid-state batteries. Nature598, 590–596 (2021). [DOI] [PubMed] [Google Scholar]
  • 3.Zhang, G. et al. A monofluoride ether-based electrolyte solution for fast-charging and low-temperature non-aqueous lithium metal batteries. Nat. Commun.14, 1081 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Zhao, X. et al. Stable binder boosting sulfide solid electrolyte thin membrane for all-solid-state lithium batteries. Nat. Energy5, 0074 (2024). [Google Scholar]
  • 5.Song, Z. et al. A reflection on polymer electrolytes for solid-state lithium metal batteries. Nat. Commun.14, 4884 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Tan, D. H. S., Banerjee, A., Chen, Z. & Meng, Y. S. From nanoscale interface characterization to sustainable energy storage using all-solid-state batteries. Nat. Nanotechnol.15, 170–180 (2020). [DOI] [PubMed] [Google Scholar]
  • 7.Zhao, Q., Liu, X., Stalin, S., Khan, K. & Archer, L. A. Solid-state polymer electrolytes with in-built fast interfacial transport for secondary lithium batteries. Nat. Energy4, 365–373 (2019). [Google Scholar]
  • 8.Su, H. et al. A scalable Li-Al-Cl stratified structure for stable all-solid-state lithium metal batteries. Nat. Commun.15, 4202 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Hashimoto, K. et al. Strain-induced crystallization and phase separation used for fabricating a tough and stiff slide-ring solid polymer electrolyte. Sci. Adv.9, eadi8505 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Shi, Z. et al. Mechanically interlocked interphase with energy dissipation and fast Li-ion transport for high-capacity lithium metal batteries. Adv. Mater.36, 2401711 (2024). [DOI] [PubMed] [Google Scholar]
  • 11.Kalnaus, S., Dudney, N. J., Westover, A. S., Herbert, E. & Hackney, S. Solid-state batteries: the critical role of mechanics. Science.381, eabg5998 (2023). [DOI] [PubMed] [Google Scholar]
  • 12.Gao, Y. & Zhang, B. Probing the mechanically stable solid electrolyte interphase and the implications in design strategies. Adv. Mater.35, 2205421 (2023). [DOI] [PubMed] [Google Scholar]
  • 13.Kim, M. S. et al. Langmuir-Blodgett artificial solid-electrolyte interphases for practical lithium metal batteries. Nat. Energy3, 889–898 (2018). [Google Scholar]
  • 14.Liu, Y. et al. Horizontal stress release for protuberance-free li metal anode. Adv. Funct. Mater. 30, 2002522 (2020).
  • 15.Wang, X. et al. Stress-driven lithium dendrite growth mechanism and dendrite mitigation by electroplating on soft substrates. Nat. Energy3, 227–235 (2018). [Google Scholar]
  • 16.Zhang, X. et al. Long cycling life solid-state Li metal batteries with stress self-adapted Li/garnet interface. Nano Lett.20, 2871–2878 (2020). [DOI] [PubMed] [Google Scholar]
  • 17.Qiu, J. et al. Enabling stable cycling of 4.2 V high-voltage all-solid-state batteries with PEO-based solid electrolyte. Adv. Funct. Mater.30, 1909392 (2020). [Google Scholar]
  • 18.Wang, H. et al. Thiol-branched solid polymer electrolyte featuring high strength, toughness, and lithium ionic conductivity for lithium-metal batteries. Adv. Mater.32, e2001259 (2020). [DOI] [PubMed] [Google Scholar]
  • 19.Li, A. et al. Nacre-inspired composite electrolytes for load-bearing solid-state lithium-metal batteries. Adv. Mater.32, 1905517 (2020). [DOI] [PubMed] [Google Scholar]
  • 20.Bai, J. et al. Limitations and strategies toward high-performance red phosphorus materials for Li/Na-ion batteries. Energy Mater. Adv.5, 0086 (2024). [Google Scholar]
  • 21.Zhang, L. et al. Lithium whisker growth and stress generation in an in situ atomic force microscope–environmental transmission electron microscope set-up. Nat. Nanotechnol.15, 94–98 (2020). [DOI] [PubMed] [Google Scholar]
  • 22.Han, S. Y. et al. Stress evolution during cycling of alloy-anode solid-state batteries. Joule5, 2450–2465 (2021). [Google Scholar]
  • 23.Hong, T. H. et al. Multiphysics factor facilitating fast reaction kinetics for high-power-density Li–Se batteries. J. Alloys Compd.948, 169789 (2023). [Google Scholar]
  • 24.Pei, A., Zheng, G., Shi, F., Li, Y. & Cui, Y. Nanoscale nucleation and growth of electrodeposited lithium metal. Nano Lett.17, 1132–1139 (2017). [DOI] [PubMed] [Google Scholar]
  • 25.Shen, X., Zhang, R., Shi, P., Chen, X. & Zhang, Q. How does external pressure shape Li dendrites in Li metal batteries? Adv. Energy Mater. 11, 2003416 (2021).
  • 26.Lin, D. et al. A silica-aerogel-reinforced composite polymer electrolyte with high ionic conductivity and high modulus. Adv. Mater. 30, 1802661 (2018). [DOI] [PubMed]
  • 27.Sun, W. et al. Amorphous FeSnOx nanosheets with hierarchical vacancies for room-temperature sodium-sulfur batteries. Angew. Chem. Int. Ed. 63, e202404816 (2024). [DOI] [PubMed]
  • 28.Han, S. et al. Sequencing polymers to enable solid-state lithium batteries. Nat. Mater.22, 1515–1522 (2023). [DOI] [PubMed] [Google Scholar]
  • 29.Liu, M. et al. Improving Li-ion interfacial transport in hybrid solid electrolytes. Nat. Nanotechnol.17, 959–967 (2022). [DOI] [PubMed] [Google Scholar]
  • 30.Fang, C.-D. et al. Revealing and reconstructing the 3D Li-ion transportation network for superionic poly(ethylene) oxide conductor. Nat. Commun.15, 6781 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Zhang, X. et al. Pressure and polymer selections for solid-state batteries investigated with high-throughput simulations. Cell Rep. Phys. Sci.4, 101328 (2023). [Google Scholar]
  • 32.Chang, C. et al. Achieving stable lithium anodes through leveraging inevitable stress variations via adaptive piezoelectric effect. Adv. Mater.36, e2313525 (2024). [DOI] [PubMed] [Google Scholar]
  • 33.Zou, P. et al. Polymorph evolution mechanisms and regulation strategies of lithium metal anode under multiphysical fields. Chem. Rev.121, 5986–6056 (2021). [DOI] [PubMed] [Google Scholar]
  • 34.Zhao, Y., Wang, R. & Martínez-Pañeda, E. A phase field electro-chemo-mechanical formulation for predicting void evolution at the Li-electrolyte interface in all-solid-state batteries. J. Mech. Phys. Solids167, 104999 (2022). [Google Scholar]
  • 35.Sun, Z.-T. et al. Real 2D galvanostatic model: encoding physicochemical heterogeneity into a full battery. Phys. Rev. Lett.135, 068001 (2025). [DOI] [PubMed] [Google Scholar]
  • 36.Yurkiv, V., Foroozan, T., Ramasubramanian, A., Shahbazian-Yassar, R. & Mashayek, F. Phase-field modeling of solid electrolyte interface (SEI) influence on Li dendritic behavior. Electrochim. Acta265, 609–619 (2018). [Google Scholar]
  • 37.Chen, L. et al. Modulation of dendritic patterns during electrodeposition: a nonlinear phase-field model. J. Power Sources300, 376–385 (2015). [Google Scholar]
  • 38.Chen, K. et al. Graphene oxide bulk material reinforced by heterophase platelets with multiscale interface crosslinking. Nat. Mater.21, 1121–1129 (2022). [DOI] [PubMed] [Google Scholar]
  • 39.Awazu, K. & Kawazoe, H. Strained Si-O-Si bonds in amorphous SiO2 materials: a family member of active centers in radio, photo, and chemical responses. J. Appl. Phys.94, 6243–6262 (2003). [Google Scholar]
  • 40.Foo, C. et al. Characterisation of oxygen defects and nitrogen impurities in TiO2 photocatalysts using variable-temperature X-ray powder diffraction. Nat. Commun.12, 661 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Lee, C., Wei, X., Kysar, J. W. & Hone, J. Measurement of the elastic properties and intrinsic strength of monolayer graphene. Science321, 385–388 (2008). [DOI] [PubMed] [Google Scholar]
  • 42.Long, H. et al. Nanomechanical properties of CH4-containing coal during CO2 storage under different injection pressures based on molecule dynamics. Appl. Surf. Sci.590, 153126 (2022). [Google Scholar]
  • 43.Wei, Y. et al. Enabling all-solid-state Li metal batteries operated at 30 °C by molecular regulation of polymer electrolyte. Adv. Energy Mater. 13, 2203547 (2023).
  • 44.Sarkar, P., Pant, P. & Nanavati, H. Physical phenomena during nanoindentation deformation of amorphous glassy polymers. Polymer312, 127597 (2024). [Google Scholar]
  • 45.Chen, K. et al. Amorphous alumina nanosheets/polylactic acid artificial nacre. Matter1, 1385–1398 (2019). [Google Scholar]
  • 46.Hu, J. et al. Dual fluorination of polymer electrolyte and conversion-type cathode for high-capacity all-solid-state lithium metal batteries. Nat. Commun.13, 7914 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Lian, H. et al. High ionic conductivity motivated by multiple ion-transport channels in 2D MOF-based lithium solid state battery. Adv. Funct. Mater. 33, 2306060 (2023).
  • 48.Wu, N. et al. Fast Li+ conduction mechanism and interfacial chemistry of a NASICON/polymer composite electrolyte. J. Am. Chem. Soc.142, 2497–2505 (2020). [DOI] [PubMed] [Google Scholar]
  • 49.Xu, S. et al. Homogeneous and fast ion conduction of PEO-based solid-state electrolyte at low temperature. Adv. Funct. Mater. 30, 2007172 (2020).
  • 50.Chen, N. et al. Biomimetic ant-nest ionogel electrolyte boosts the performance of dendrite-free lithium batteries. Energy Environ. Sci.10, 1660–1667 (2017). [Google Scholar]
  • 51.Zhou, X. et al. Difunctional MOF for dendrite-free all-solid-state lithium metal batteries by synergistic effect of hydrogen bond and electrostatic interaction. Nano Energy108, 108221 (2023). [Google Scholar]
  • 52.Yang, W. et al. Local electronic structure modulation of interfacial oxygen vacancies promotes the oxygen activation capacity of Pt/Ce1-xMxO2-δ. ACS Catal.14, 5936–5948 (2024). [Google Scholar]
  • 53.Luo, B. et al. Facilitating ionic conductivity and interfacial stability via oxygen vacancies-enriched TiO2 microrods for composite polymer electrolytes. Chem. Eng. J.460, 141329 (2023). [Google Scholar]
  • 54.Ruan, J. et al. Enhancing the whole migration kinetics of Na+ in the anode side for advanced ultralow temperature sodium-ion hybrid capacitor. J. Power Sources13, 2301509 (2023). [Google Scholar]
  • 55.Sun, J. et al. Facilitating interfacial stability via bilayer heterostructure solid electrolyte toward high-energy, safe and adaptable lithium batteries. Adv. Energy Mater.10, 2000709 (2020). [Google Scholar]
  • 56.Xu, C. et al. Interface layer formation in solid polymer electrolyte lithium batteries: an XPS study. J. Mater. Chem. A2, 7256–7264 (2014). [Google Scholar]
  • 57.Ismail, I., Noda, A., Nishimoto, A. & Watanabe, M. XPS study of lithium surface after contact with lithium-salt doped polymer electrolytes. Electrochim. Acta46, 1595–1603 (2001). [Google Scholar]
  • 58.Yu, W., Yu, Z., Cui, Y. & Bao, Z. Degradation and speciation of Li salts during XPS analysis for battery research. ACS Energy Lett.7, 3270–3275 (2022). [Google Scholar]
  • 59.Lahiri, A., Borisenko, N., Borodin, A., Olschewski, M. & Endres, F. Characterisation of the solid electrolyte interface during lithiation/delithiation of germanium in an ionic liquid. Phys. Chem. Chem. Phys.18, 5630–5637 (2016). [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

41467_2026_78129_MOESM2_ESM.pdf (181.6KB, pdf)

Description of Additional Supplementary Files

Supplementary Data 1–7 (42.9KB, zip)
Peer Review File (7.3MB, pdf)
Source Data (28.4MB, xlsx)

Data Availability Statement

The data that support the findings and conclusions generated in this study are provided in the main article, the Supplementary Information, and the Supplementary Data. Source data are provided with this paper.


Articles from Nature Communications are provided here courtesy of Nature Publishing Group

RESOURCES