Swelling-induced strain-isolation for stretchable and breathable integrated electronics

Abstract

Wearable electronics hold great promise for health management and chronic disease monitoring, yet their widespread applications are limited by insufficient stretchability, breathability, and mechano-electrical stability. Herein, we report a swelling-induced modulus design that enables the construction of breathable and strain-isolated substrates for integrated electronics. The swelling-induced in-situ polymerization of acrylates within polyurethane membrane effectively enhances the mechanical properties while preserving air permeability. In addition, by optimizing substrate modulus and interfacial contact, rigid islands for supporting functional devices and soft bridges for assembling interconnects are rationally designed. As a result, the circuit board achieves high stretchability (519%, 3.86 MPa) and electrical stability (ΔR/R₀ = 3.78 under 500% strain), together with high air (31.4 mm s⁻¹) and moisture (2577.67 g m⁻² day⁻¹) permeability, outperforming previously reported stretchable and breathable supports. A stretchable and breathable system is also demonstrated, which integrates wireless charging, energy storage, and sensing for posture monitoring with long-term comfortable wear.

A breathable strain-isolated stretchable substrate is developed through a swelling-induced modulus design, which facilitates the integration of wireless charging, energy storage, and sensing units into a wireless wearable posture-monitoring system.

Introduction

Wearable electronic systems, by integrating multifunctional components for signal sensing, collection, and transmission, are crucial for future physiological monitoring, health management, and chronic disease care^1–5. Since diverse multifunctional modules with different functionalities should be integrated onto a supporting substrate, its material characteristics and mechanical properties critically determine the device durability, stability, and user comfort^6–8. Conventional electronic systems typically employed rigid printed circuit boards (PCB) as supporting substrates. Although they offer high electrical performance and long-term stability, the lack of flexibility and mechanical compliance limits their integration with soft and deformable human tissues^9–12. With this concern, flexible electronic systems based on elastic substrates have been developed. For example, Ecoflex and polydimethylsiloxane (PDMS) could provide mechanical flexibility and conformal skin contact¹³. However, these dense polymer substrates suffer from low breathability and poor mechanical tolerance, leading to skin discomfort, interfacial degradation, and reduced reliability during prolonged wear¹⁴. Therefore, it is essential to develop novel substrates for stretchable and breathable electronics to improve wearer comfort, device stability, and multifunctional integration^15,16.

Porous interconnected fibrous materials, such as electrospun membranes and textiles, offer excellent mechanical robustness with efficient air and moisture transport, which are promising substrates for stretchable and breathable electronics^17–20. For instance, Cao et al. embedded a liquid metal-polyvinyl alcohol (LM-PVA) conductive network into an electrospun Poly(styrene-block-butadiene-block-styrene) membrane, forming a stretchable circuit even at 100% strain²¹. Peng et al. deposited functional layers on a styrene-ethylene-butylene-styrene (SEBS) electrospun membrane, successfully maintaining superior breathability of 1439.1 gm⁻² day⁻¹²². Although these substrates are well suited for stretchable and breathable systems with soft functional components, integrating rigid functional units (e.g., power, sensing, and transmission modules) while maintaining structural and electrical stability remains challenging²³. During skin deformation, huge stress concentrates at the interface between the rigid units and the soft substrate, causing delamination and performance failures²⁴.

To alleviate the stress concentrations for stretchable and breathable electronics, researchers have employed modulus engineering to pattern fibrous membranes with rigid and soft regions^25–28. For instance, Flavin et al. used mechanical clips²⁹ and Wicaksono et al. employed commercial encapsulants³⁰ to create locally stiff regions on membranes, where rigid islands carry hard components and soft regions enable circuit interconnections, achieving integration of multifunctional device. However, such physically imposed stiffening introduces abrupt modulus discontinuities, causing severe stress concentration and interfacial failure, and the encapsulation layers also drastically reduce the intrinsic breathability of the electronics. Material modifications have also been explored to locally enhance stiffness. For example, Wang et al. infiltrated PVA into porous polyurethane (PU) films to form a composite substrate with locally reinforced modulus³¹. Qu et al. coated electrospun SEBS substrates with PU to form locally stiffened islands, where inter fiber interactions at the interface effectively redistributed stress and alleviated local stress concentration, achieving strain-isolation³². Although these methods effectively regulate local modulus, their reinforcement primarily originates from two-phase interfaces, where mechanical coupling is largely localized at the interfacial region. In addition, filling or coating substrates with a stiff secondary phase would block pores and severely disrupt gas transport and reduce breathability³³. The limitations of the reported stretchable and breathable substrates can be originated from the heterogeneous pore distribution and local modulus fluctuations within porous substrates, which endow it challenging to achieve spatial modulation of modulus, robust intermolecular coupling, and preserved air permeability within the same matrix³¹. Therefore, novel design strategies are still urgently required to develop multifunctional electronics with high stretchability, breathability, and mechano-electrical stability.

Herein, we report a swelling-induced modulus design for constructing breathable strain-isolated stretchable substrate (SISS) for highly integrated electronics, which employs monomer swelling and in-situ polymerization to reconstruct individual fibers and generate an interpenetrated network. Leveraging strong intermolecular interactions, PU nanofibers are swollen by isobornyl acrylate/butyl acrylate (IBOA/BA) and subsequently undergo in-situ photopolymerization. The formed polyacrylate (PA) not only achieves mechanical reinforcement through chain entanglement and hydrogen bonding throughout the reinforced fibers but also preserves the porous structure and intrinsic breathability of the PU membrane. In addition, by tuning the ratio of rigid and soft segments (IBOA/BA) in the PA and controlling their spatial distribution within the PU substrate (PUS), the mechanical heterogeneity and structural stability can be well maintained, which facilitates the reliable mechano-electrical performance. As a result, the obtained SISS achieves high stretchability (519%, 3.86 MPa) and electrical stability (ΔR/R₀ = 3.78 under 500% strain), together with high air (31.4 mm s⁻¹) and moisture (2577.67 g m⁻² day⁻¹) permeability. The programmable SISS enables the reliable integration of wireless charging, energy storage, and sensing into a battery-integrated wireless device, demonstrating stable operation during continuous posture monitoring in bedridden individuals.

Results

Concept of swelling-induced modulus design

Conventional electronics based on rigid PCB exhibit poor mechanical compliance with soft tissues and are intrinsically non-breathable, severely limiting their conformability and wearability. Although porous interconnected fibrous electrospun membranes offer high flexibility and breathability, the direct integration of rigid functional components induces severe interfacial stress concentration under deformation, resulting in device delamination and unstable electrical performance. Previously reported modulus engineering to pattern fibrous membranes with rigid and soft regions enable multifunctional integration but suffer from abrupt modulus mismatch, localized interfacial interactions, and compromised breathability (Fig. 1a). In this work, as shown in Fig. 1b, a stretchable and breathable electrospun PU nanofiber membrane was employed as the pristine substrate. Without sacrificing its porous and breathable structure, a swelling-induced modulus design was employed to generate a programmable rigid-island/soft-bridge architecture. Rigid functional components are distributed on rigid-island regions, while stretchable circuits are arranged along the soft-bridge regions. This strategy effectively addresses the limitations of conventional platforms, including insufficient stretchability, poor breathability, and limited mechano-electrical stability, thereby improving wearing comfort and practical usability for integrated wearable electronics.

Fig. 1. Swelling-induced strain-isolation for stretchable and breathable integrated electronics.

a Schematic of the previously reported integrated electronics under tensile strain. b Exploded schematic of a typical SISS-based electronics, consisting of a stretchable and breathable SISS, stretchable circuits, and rigid electronic components. c Schematic of swelling-induced modulus design and the deformation of PU, PU/Stiff-PA, and PU/Soft-PA fibers under tensile strain. d Magnified schematics highlighting components on rigid islands and stretchable circuits on soft bridges of the SISS in the initial and stretched states. e Schematic of a stretchable and breathable battery-integrated wireless sensing system based on the SISS, integrating a wireless charging module, energy storage module, and a piezoresistive sensing module.

The concept of swelling-induced modulus design and the preparation of SISS are illustrated in Fig. 1c. Different from conventional local reinforcement strategies, where mechanical enhancement is primarily localized at heterogeneous interfaces, this strategy reconstructs individual fibers through monomer infiltration and in-situ polymerization, resulting in reinforcement throughout the modified fibrous network. Benefiting from the strong molecular affinity between acrylate monomers (IBOA/BA) and electrospun PU nanofibers, the monomer molecules infiltrate the PU matrix and induce fiber swelling. Subsequent in-situ photopolymerization generates PA within the fibers. The resulting PA chains form entanglements and hydrogen bonding interactions with PU molecular chains, significantly enhancing load transfer and stress sharing within individual fibers, thereby improving their mechanical robustness and resistance to fracture under tensile deformation. Furthermore, by controlling the molecular composition of the PA phase, the local modulus of the fibers can be spatially regulated. The rigid cyclic structure of IBOA introduces enhanced chain constraint and stiffness in rigid regions, whereas the compliant linear BA segments promote strain accommodation in soft regions. Importantly, as monomer swelling and polymerization occur within individual fibers, the original fibrous morphology and interconnected porous network are preserved, ensuring continuous gas transport pathways and maintaining superior breathability. This molecular modulation enables programmable spatial distribution of soft and rigid regions within the breathable fibrous substrate (Fig. 1d). During repeated stretching, the rigid-island regions provide mechanical support and interfacial constraint, preventing delamination and detachment of functional components, while the soft-bridge areas accommodate the dominant deformation and preserve continuous conductive pathways. Moreover, constructing the rigid and soft regions from the same composite system through a tunable IBOA/BA ratio greatly eliminates interfacial mismatch, thereby ensuring robust interfacial integrity and stable mechano-electrical performance during long-term operation.

This work integrates material design with structural engineering through swelling-induced modulus design, thereby bridging molecular-scale fiber reinforcement with macroscale device strain-isolation behavior and mechanical reliability, enabling the SISS with high stretchability, breathability, and mechano-electrical stability. On this basis, the SISS further enables programmable integration and structural layout of multifunctional modules. For example, by incorporating wireless charging module, energy storage, and piezoresistive sensing units within the rigid islands and interconnections on the soft bridges, a battery-integrated wireless sensing system is assembled. This system demonstrates promising application potential in stretchable, breathable, and comfortable wearable electronics for physiological signal monitoring, providing a representative design paradigm for multifunctional electronic systems (Fig. 1e).

Preparation and characterization

As shown in Fig. 2a, the SISS was fabricated by programmably printing modulus-tuned polyacrylate precursors (PAp) onto electrospun PU membranes (detailed molecular structures in Supplementary Fig. 1), followed by in situ photopolymerization. During printing, acrylate monomers infiltrate the interchain space of PU and disrupt hydrogen bonding due to the matched polarity and solubility parameters between acrylate esters and PU. This leads to the disentanglement of soft PU segments (S-PU), reducing the local chain packing density and inducing fiber swelling, while the hard segments (H-PU) remain largely intact³⁴. The swelling process generates additional free volume and capillary pathways, which promote monomer penetration into the fiber interior rather than surface accumulation. Subsequent in-situ photopolymerization locks the infiltrated monomers within the swollen fibers, yielding a PU/PA interpenetrating double network structure. Meanwhile, the PU-PA hydrogen bonds are established between the –NH groups of PU and the C=O groups of PA (Fig. 2b). The strain-isolation design can be achieved by tuning the ratio of rigid IBOA to flexible BA in the PAp network.

Fig. 2. Preparation and characterization of the SISS.

a Schematic of the fabrication process and microstructures of the SISS. b Detailed chemical structure of crosslinked PU/PA. c XRD patterns and d FTIR spectra of the PU, PU/PAp, and PU/PA. e Photographs of PU, PU/PAp, and PU/PA membranes (Soft/PA on the top and Stiff/PA on the bottom). f SEM image of the PU membrane with in situ polymerized Soft-PA (left) and Stiff-PA (right). High-magnification views of the g PU/Soft-PA and h Stiff-PA/PU interface. The experiment was conducted three times with similar results. i Photographs of the PU membrane stitched by Soft-PA and the stitched PU under stretching. j Statistical comparison of the DFT calculated binding energies of PU with IBOA, BA, and dimethylsiloxane. Electrostatic potential mapping of k PU with IBOA and l PU with BA. The molecular structures and electrostatic potential maps of the IBOA and BA molecules are presented below.

To verify the swelling behavior of Soft/Stiff-PAp on the PU membrane, various characterizations were conducted, with PDMS and Ecoflex precursors (PDMSp and Ecoflexp) serving as control materials. As depicted in Supplementary Fig. 2, the contact angles between Soft/Stiff-PAp and PU are as low as 25.5° and 32.0°, respectively, much smaller than those of PDMSp (66.0°) and Ecoflexp (59.8°), indicating excellent wettability. XRD patterns reveal that the diffraction peaks of PU/Soft-PAp and PU/Stiff-PAp are significantly left-shifted compared to pristine PU, indicating an ordered increase in d-spacing resulting from swelling induced chain expansion. After in-situ photopolymerization, the peak remains at the shifted position, suggesting that the swollen configuration is effectively preserved and topologically constrained within the interpenetrating PU/PA network (Fig. 2c). In contrast, the diffraction peaks of PU/PDMS and PU/Ecoflex show no noticeable shift, indicating negligible swelling (Supplementary Fig. 3). The fourier-transform infrared spectroscopy (FTIR) results confirm the swelling of PU and the formation of the PU/PA network (Fig. 2d). Upon PAp infiltration, the C-O stretching peak of PU at ~1073 cm⁻¹ red shifts to ~1054 cm⁻¹, indicating a more relaxed local environment due to hydrogen bonding and solvation interactions between the acrylate monomer and PU chains. After polymerization, the peak shifts back to higher wavenumbers, implying reduced chain mobility as PU chains become immobilized within the interpenetrating PU/PA network. In addition, the vinyl group peaks in PAp (~974 and 809 cm⁻¹) disappear after polymerization, indicating the conversion of acrylate C=C bonds into σ-bonded polymer chains and the formation of the crosslinked PU/PA network. Consistent with XRD and FTIR, PU membranes become transparent after PAp monomer infiltration, indicating the occurrence of swelling and the penetration of monomers into the fiber interior (Fig. 2e). In contrast, PU membranes treated with PDMSp and Ecoflexp remain white and opaque, suggesting that no appreciable swelling occurs (Supplementary Fig. 4). Mechanical testing was conducted to further verify the swelling behavior. Cyclic tensile tests show that PU/Soft-PAp and PU/Stiff-PAp cannot fully recover after 200% strain, exhibiting a residual strain of ~40%, whereas pristine PU and polymerized PU/Soft-PA fully recover their original shapes (Supplementary Fig. 5). This behavior is consistent with monomer induced swelling and temporary softening of the PU matrix, whereas polymerization locks the chains into the PU/PA network and recovers the intrinsic elastic response. To quantitatively evaluate the swelling behavior, the volume expansion ratio of the PU membrane was measured at different PAp loadings. As shown in Supplementary Fig. 6 and Fig. 7, the volume expansion ratio increases from 73% for PU/PA₂ to 124% for PU/PA₄ (x in Soft/Stiff-PA_x denotes the mass ratio of PU:PA = 1:x), demonstrating that the swelling degree can be systematically controlled through precursor content.

Scanning electron microscope (SEM) images (Fig. 2f–h) reveal a well-integrated interface between PU and Soft/Stiff-PA, showing continuous interfacial contact and a gradual transition while the intrinsic porous architecture of the PU membrane remains largely preserved. Raman mapping (Supplementary Fig. 8) and atomic force microscopy (AFM) nanomechanical characterization (Supplementary Fig. 9) confirm the existence of a well-defined gradient interface between the PU and PU/PA domains, with a transition width of approximately 40 μm. To evaluate interfacial robustness, red dyed PU was stitched to PU using Soft-PA. The stitched sample remained intact under 500% tensile strain, demonstrating strong interfacial adhesion (Fig. 2i). Density functional theory (DFT) calculations were conducted to investigate the molecular interactions underlying the swelling behavior. Binding interactions between IBOA, BA, and dimethylsiloxane (PDMS monomer) with PU were evaluated at two distinct adsorption sites. The binding energies between PU and IBOA (−3.6 and −4.8 kcal mol^-1) and between PU and BA (−8.8 and −5.2 kcal mol⁻¹) are significantly higher than those between PU and dimethylsiloxane (−2.1 and −1.1 kcal mol⁻¹) (Fig. 2j and Supplementary Fig. 10), indicating stronger PU-IBOA and PU-BA interactions that support the swelling process and the enhanced mechanical strength of the PU/PA system. Analysis of the optimized geometries indicates that the enhanced PU-IBOA and PU-BA interactions arise from multiple factors. Electrostatic interactions contribute significantly to the enhanced binding, as evidenced by electrostatic potential calculations performed at two distinct adsorption sites (Fig. 2k, l), which reveal clear charge complementarity between the urethane groups of PU and the ester groups of IBOA and BA (Supplementary Fig. 11). In addition, hydrogen bonding between the -NH groups of PU and the carbonyl oxygen atoms of IBOA and BA further strengthen the intermolecular interactions, leading to the higher binding energies of the PU-IBOA and PU-BA systems. Together with additional van der Waals contributions, these interactions collectively stabilize the PU-monomer complexes, lowering the free energy of system and promoting monomer diffusion into the PU network.

Mechano-electrical performance

The SISS with a rigid-island/soft-bridge architecture was evaluated for mechanical robustness and mechano-electrical stability under deformation. To optimize the performance of the SISS, the effect of PA content on the mechanical properties was first investigated. As shown in Fig. 3a, the elongation at break of PU exhibits a slight increase with increasing Soft-PA content, while the tensile strength significantly increases and plateaus at 3.74 MPa for Soft-PA₃, 52.0% higher than pristine PU. In contrast, increasing the Stiff-PA content leads to a rapid reduction in elongation after polymerization, while the tensile strength increases quickly and approaches saturation, reaching 29.06 MPa for Stiff-PA₃ (Fig. 3b). To achieve an optimized rigid-island/soft-bridge architecture, Stiff-PA₃ and Soft-PA₃ were selected for their favorable balance of substrate mechanical properties and permeability (next Section) (unless otherwise specified below, Soft/Stiff-PA refers to Soft/Stiff-PA₃). The Young’s modulus (E) of PU/Stiff-PA for rigid-island is measured to be 1066.13 MPa, while that of PU/Soft-PA for soft-bridge is 0.60 MPa, resulting in a modulus ratio (ΔE = E_rigid/E_soft) of 1777 (Supplementary Fig. 12a). Notably, PU/Stiff-PA and PU/Soft-PA exhibit higher tensile strength than Stiff-PA and Soft-PA, respectively, benefiting from the co-network formation and strong interfacial bonding between PA and PU (Supplementary Fig. 12b). In addition, PU/Soft-PA also shows higher mechanical strength than the control samples PU/PDMS and PU/Ecoflex (Supplementary Fig. 12c).

Fig. 3. Mechano-electrical performance of the SISS.

Stress and strain values of PU with varying a Soft-PA and b Stiff-PA content. c Stress-strain curves of PUS and SISS. d FEA simulation of the stress distribution in PUS, PU-Stiff island, and SISS under tensile deformation. e Enlarged views of the interfacial stress evolution during stretching. f ΔR/R₀ of a 1 Ω resistor on the rigid island with LM interconnects under strain. g Resistance changes in cycled stretching test of a 0 Ω resistor on the rigid island with LM interconnects. Inset: enlarged view of the 2000th cycle and the corresponding photographs, in which the resistor is replaced by an LED under stretching. h Resistance changes of resistors on rigid islands connected by LM interconnects under tensile strain. i Comparison of modulus ratios (ΔE) and ΔR/R₀ versus limiting strain (ε_f) between this work and previously reported substrates. j Resistance of seven 10 Ω resistors in parallel on a rigid island with LM interconnects under strain. Inset: Photographs of the parallel LED array under twisting and stretching. Scale bars: 1 cm.

The fabricated SISS geometry featuring Soft-PA bridges that connect two Stiff-PA islands was employed to investigate mechanical performance. As shown in Fig. 3c and Supplementary Fig. 13, strain isolation introduced by non-stretchable rigid islands leads to a slight reduction in overall elongation compared with PUS, which nevertheless remains high at 519%, while the tensile strength increases by 56.9% to 3.86 MPa. Finite element analysis (FEA) was further employed to visualize the stress distribution of rigid devices integrated on the PUS, PUS-Stiff island, and SISS (Supplementary Fig. 14). For the PUS, FEA reveals significant stress concentration at the electronic component pins (Fig. 3d, upper) due to the modulus mismatch between the rigid device and the soft substrate, indicating a high risk of device detachment. Although introducing a PUS-Stiff island beneath the electronic component alleviates stress within the device region, stress localization still occurs at the interface of PUS-Stiff island due to the heterogeneity of the interfacial materials and the abrupt modulus mismatch between Stiff-PA and PU (Fig. 3d, middle). In contrast, the SISS exhibits a stable stress distribution, with the rigid island region largely free of stress and only small stress observed at the interfaces between Stiff-PA, Soft-PA, and PU (Fig. 3d, bottom). This behavior can be attributed to the use of homogeneous materials for the soft and stiff regions combined with a modulus gradient design, in which the Soft-PA acts as a buffer between the Stiff-PA and the PU, smoothing strain transfer across interfaces and suppressing stress concentration. Moreover, under a wide strain range (0–500%), the interfacial stress distribution of the SISS structure remains nearly unchanged, in sharp contrast to the PU and PU-Stiff island substrates, indicating a dynamically adaptive interface that preserves mechanical integrity under large deformation (Fig. 3e). In-situ optical micrographs of the PU/Soft-PA and PU/Stiff-PA interface under stretching up to 500% also confirm that deformation is localized in the PU/Soft-PA, while the PU/Stiff-PA remains undeformed, with no interfacial cracking observed (Supplementary Fig. 15).

The electrical performance stability of devices integrated on SISS was then measured. As a proof of concept, a 1 Ω resistor was affixed to a rigid island and connected via liquid metal (LM) interconnects (Supplementary Fig. 16) across a soft bridge. Benefiting from its strong interfacial adhesion, the PU/Soft-PA substrate enables stable interfacial integration of soft interconnects and electronic components (interfacial fracture energy as high as 571.4 J m⁻², more than an order of magnitude higher than those of control substrates, including PU, PU/Ecoflex, and PU/PDMS; all <40 J m⁻², Supplementary Fig. 17). As shown in Fig. 3f, the circuit remains functional under 500% strain with only a minor relative resistance change (ΔR/R₀ = 3.78). In contrast, an identical interconnect on PUS fails at 39.4% strain, accompanied by a pronounced resistance increase. Under 100% strain for more than 10,000 cycles, the LM interconnects in the SISS remain electrically stable with only minimal resistance variation (Fig. 3g). An inset further visualizes that the LED mounted on the rigid island remains lit even at 500% strain. When connected to different resistors ranging from 100 to 400 Ω, the resistance of the circuit shows negligible change when stretched to 500% strain (Fig. 3h). Importantly, as shown in Fig. 3i, the designed SISS combines excellent stretchability with a high modulus contrast (ΔE), outperforming previously reported heterogeneous and homogeneous island-bridge designs (Supplementary Table 1)^{25,28,35–41}. With a high ultimate strain of 500% and a low ΔR/R₀ of 3.78, it also surpasses earlier systems in both stretchability and electrical stability (Supplementary Table 2)^2,30,42–49. To further evaluate mechano-electrical stability at the array level, seven resistors were paralleled and arranged in a hexagonal geometry with a 20 mm centre-to-centre spacing (Fig. 3j). The resulting parallel array (~1.5 Ω), together with individual resistors (10 Ω), exhibits negligible resistance change even at 500% tensile strain. Replacing the resistors with LEDs produced a stretchable LED array that performs robustly under diverse deformation modes, including torsion and stretching (inset of Fig. 3j and Supplementary Video 1).

Wearing comfort

The effect of PA content on breathability was first tested. As shown in Fig. 4a and Supplementary Fig. 18, at low PA contents, the PU membrane retains an open fibrous porous structure, but as the content exceeds PA₃, inter-fiber voids fill up, and air permeability decreases accordingly. Nevertheless, Soft-PA₃ and Stiff-PA₃ retain sufficient permeability (>20 mm s⁻¹), remaining comparably high relative to pristine PU (~35.6% and 39.8% decrease), ensuring adequate gas transport while providing mechanical reinforcement, thus were selected as the target samples (Fig. 4b). In contrast, Ecoflex and PDMS form dense layers over the PUS, eliminating its porous structure and greatly reducing gas transport (Supplementary Fig. 19). Only PU/PDMS_2.5 exhibits weak permeability (1.73 mm s⁻¹), while PU/Ecoflex is essentially impermeable, demonstrating that such simple coating strategy fails to preserve air permeability.

Fig. 4. Wearing comfort of SISS.

a Schematic and SEM images of PU membranes with varying PA contents. The experiment was conducted three times with similar results. b Comparison of air permeability among PU membrane with varying PA, Ecoflex, and PDMS contents. c Air permeability, d moisture permeability, e WVTR, f hydrostatic pressure, g stiffness, and h abrasion resistance of PUS, SISS, PPAS, and PPS. Bar height, mean; error bars, standard deviation (s.d.); n = 5 independent samples. i Participant feedback on the comfortable wearing duration of different electronics during a 12-h test. j Comparison of relative skin moisture content after 6 h of wearing PUS, SISS, PPAS, and PPS. Square, mean; center line, median; box limits, upper and lower quartiles; whiskers, 1.5 × interquartile range; points, outliers. n = 10 volunteers. Statistical significance was assessed using two-sided multiple-comparisons tests. k Photographs showing the electronic devices based on PUS, SISS, PPAS, and PPS attached to the skin (left) and the corresponding skin condition after 6 h of wear (middle). Infrared thermal images (right) show the skin temperature after removal. l Two glass bottles containing HCl and NH₃·H₂O covered with electronics based on PUS, SISS, PPAS, and PPS.

The wearing comfort of SISS was then evaluated in terms of air and moisture permeability, waterproofing, flexibility, and abrasion resistance, and compared with three control substrates, namely PUS, PPAS (PUS fully modified with PA), and PPS (PU/PDMS_2.5 substrate). As shown in Fig. 4c, d, the air and moisture permeabilities of SISS reach 31.4 mm s⁻¹ and 2577.67 g m⁻² day⁻¹, which are comparable to those of PUS (34.9 mm s⁻¹ and 2644.37 g m⁻² day⁻¹) and higher than PPAS (20.5 mm s⁻¹ and 2047.93 g m⁻² day⁻¹), and are 18-fold and 21-fold higher than those of PPS (1.7 mm s⁻¹ and 120.52 g m⁻² day⁻¹), respectively. Water vapor transmission rate (WVTR), which quantifies the substrate capacity to transmit sweat, shows consistent trends (Fig. 4e). The SISS shows a WVTR of 377.87 kg cm⁻² (measured over 77 h), which is comparable to that of PUS (393.34 kg cm⁻²), slightly higher than that of PPAS (324.68 kg cm⁻²), and substantially higher than that of PPS (31.03 kg cm⁻²). The high permeability of SISS can be attributed to the selective formation of PU/PA regions and their swelling behavior, which effectively maintain the inter-fiber porosity and network integrity. Hydrostatic pressure was used to determine the waterproof properties of substrates. As shown in Fig. 4f, PPS exhibits the highest hydrostatic pressure due to its dense structure, but this high hydrostatic pressure is accompanied by a sacrifice in breathability. SISS achieves a balanced combination of breathability and protection, with a hydrostatic pressure of 115.85 mbar that is lower than PPAS and PPS yet higher than PUS (103.44 mbar), remaining sufficient to meet basic barrier requirements. The superior flexibility of the SISS endows it with good conformability to dynamically deforming skin. Bending stiffness measurements indicate that SISS exhibits bending performance comparable to PUS (30.7 mm), with an elongation length of 35.6 mm, whereas PPAS (168.0 mm) and PPS (161.2 mm) are substantially stiffer due to their higher moduli and denser composite architectures (Fig. 4g). Wear resistance was also evaluated in Fig. 4h and Supplementary Fig. 20. SISS can withstand up to 360 abrasion cycles, which is lower than that of PPAS and PPS but still approximately 579% higher than PUS. Such enhanced durability originates from the strain-selective distribution of PU/PA in programmed regions, where it preferentially accommodates localized wear, achieving an optimal balance between durability and flexibility. Notably, unlike non-breathable elastomers^46,50–52 such as PDMS⁵³ or breathable yet weakly stretchable electrospun and textile substrates^{19,23,54–60}, SISS simultaneously achieves higher air/moisture permeability and mechanically robust stretchability (Supplementary Table 3).

To evaluate wearability and skin comfort, a LED module was integrated on the PUS, SISS, PPAS, and PPS as a demonstrator (Supplementary Fig. 21). In a user study with ten participants (five male and five female), four substrates were affixed to the wrist for up to 12 h. Participants were allowed to remove patches early if discomfort occurred, and wearing durations were recorded. The results indicate that PUS and SISS remain wearable for considerably longer durations than PPAS, and far longer than PPS (Fig. 4i). Relative skin-moisture measurements on the wrist were also conducted after 6 h of skin coverage with PUS, SISS, PPAS, and PPS. PPS causes a substantial increase in local skin moisture (~68%), whereas PUS and SISS maintain low moisture levels (~35 and 37%) and PPAS shows only a moderate increase (~47%), highlighting the good breathability of SISS (Fig. 4j). Furthermore, a 6 h wrist-wear test was conducted for visual evaluation of skin response (Fig. 4k). The areas covered by PUS, SISS, and PPAS remain free of inflammation and maintain skin temperature within comfortable ranges, whereas impermeable PPS induces serious erythema and heat discomfort at the wrist. A simple visual permeability test was also conducted by generating white fumes through the reaction between HCl and NH₃ in a sealed bottle. White fumes were observed from bottles covered with PUS, SISS, and PPAS, but not from those covered with PPS (Fig. 4l), in agreement with the quantitative permeability measurements. Moreover, durability under skin-contact conditions is also critical for practical applications. As the key functional component of the SISS, the PU/PA region exhibits superior stability after 3 days of artificial sweat (pH 4.7 ± 0.1) exposure, with negligible changes in its mechanical, structural, and interfacial properties (Supplementary Fig. 22 and Fig. 23).

Integrated device based on SISS

Based on the mechanically robust and spatially programmed SISS architecture, a stretchable integrated device was realized by assembling wireless charging, energy storage, and piezoresistive sensing modules on the rigid-island regions, with electrical interconnections on the soft-bridge regions. As shown in Fig. 5a, the LM antenna was printed on the rigid-island region to prevent coil deformation, which otherwise causes antenna coil mismatch and increased resistance (photograph provided in Supplementary Fig. 24). This design achieves stable wireless power reception from the transmitting coil via inductive coupling. During wireless charging, a 10 V sinusoidal alternating current (AC) voltage was applied to the transmitting coil to generate an alternating magnetic field, inducing a 6 V AC signal in the LM antenna, which was subsequently rectified into a direct current (DC) voltage of ~6 V (Fig. 5b). Within the stretching range of 0-100%, the output voltage of the LM receiving antenna (Fig. 5c) remains nearly constant at 6 V, indicating high wireless energy transfer and harvesting efficiency suitable for battery charging.

Fig. 5. High-performance integrated device based on SISS.

a Schematic illustration of wireless charging device. b Voltage waveforms applied to the transmitting coil, received by the LM antenna, and rectified for DC charging. c Voltage waveforms received by the LM antenna under 100% strain. d Schematic illustration of the ZIB. e Comparison of GCD curves at 1 mA cm⁻² for the ZIB in the initial state and at 100% strain. f GCD curves of two stretchable ZIBs connected in series. g Schematic of a piezoresistive sensor. h ΔI/I₀ of the piezoresistive sensor versus time under pressures ranging from 0.1 to 120 kPa. Inset: ΔI/I₀ change versus pressure. i ΔI/I₀ of the piezoresistive sensor in the initial state and under 100% strain. Inset: enlarged comparison of the response and recovery times. j Wireless charge and galvanostatic discharge curves of ZIB. k Real time ΔI/I₀ change response curve (blue) of the piezoresistive sensor during continuous monitoring and the corresponding discharge voltage curve (yellow) of the ZIB.

Considering the advantages of safety, cost, and environmental friendliness, zinc-ion battery (ZIB) was selected as the energy storage module⁶¹. The Zn anode and V-based cathode (SEM and XRD results in Supplementary Fig. 25) were assembled to fabricate the ZIB (Fig. 5d). From the cyclic voltammetry (CV) curves in Supplementary Fig. 26, the ZIB exhibits two pairs of cathodic/anodic peaks, corresponding to a two-step reverse oxidation/reduction reaction. In Fig. 5e, the galvanostatic charge-discharge (GCD) curves show that ZIB delivers a high capacity of 0.52 mAh cm⁻² and superior energy/power densities of 0.54 mWh cm⁻²/1.04 mW cm⁻² (107.96 Wh kg⁻¹/208.69 W kg⁻¹) at a current density of 1 mA cm⁻² and exhibits negligible change even at 100% tensile strain, confirming the excellent mechanical stretchability and electrochemical stability. The long-term stability of the ZIB was also examined in Supplementary Fig. 27, in which 89.4% capacity retention can be maintained after 250 cycles. Moreover, the stretchable ZIBs can be connected in parallel (Supplementary Fig. 28) to realize doubled discharging time (i.e., capacitance), while operating voltage is 3.2 V when two ZIBs are connected in series (Fig. 5f), exhibiting great potential for powering sensing systems.

The pressure sensor was also assembled on the rigid-island region to ensure strain-isolation and minimize strain-induced signal interference. The device consists of Cr/Ag interdigitated electrodes (Supplementary Fig. 29) and a PU/MXene porous sensing layer (Supplementary Fig. 30) laminated on top (Fig. 5g). Under external loading, the PU/MXene porous layer is compressed, which promotes nanosheet interconnection, increases the electrode contact area, and generates a measurable electrical response. Figure 5h shows that within a wide pressure range of 0.1–120 kPa, the current response increases significantly with increasing pressure, enabling real-time pressure detection. The sensor shows a sensitivity of 11.4 kPa^–1 at 0.1–1 kPa, decreasing to 4.54 and 1.89 kPa⁻¹ at 1–10 kPa and 10-120 kPa, respectively, due to the initial rapid formation and subsequent saturation of conductive pathways among MXene nanosheets. Furthermore, under an applied pressure of 1 kPa, the sensing amplitude as well as the response and recovery times (~130 ms) remain nearly unchanged between the initial and 100% stretched states (Fig. 5i), confirming the superior operational reliability of the device.

With the integration of functional modules, a battery-integrated wireless sensing device was successfully constructed. The LM antenna, ZIB, and piezoresistive sensor were electrically interconnected via LM traces printed on the soft-bridge regions, enabling seamless energy transfer from the wireless charging module to the ZIB, and from the ZIB to the sensor. As shown in Fig. 5j and Supplementary Fig. 31, the charge-discharge curves of a ZIB exhibit a distinct wireless charging plateau, reaching 1.45 V (90.6% of the full charge voltage measured by GCD), followed by a stable constant current discharge plateau over a current density range of 1–4 mA cm⁻², demonstrating the effective establishment of the wireless charging system. The wireless charging ZIB also exhibits good cycling stability, retaining over 85% of its initial discharge capacity after 30 wireless charge/discharge cycles (Supplementary Fig. 32). Furthermore, the ZIB power supply provides a stable output voltage of ~1.0 V, driving the piezoresistive sensor over 100 loading/unloading cycles. Under an applied pressure of 70 kPa, the sensor exhibits a relative current change (ΔI/I₀) of ~165 and consistently delivers reproducible sensing signals, demonstrating the reliable long-term operation of the integrated system (Fig. 5k).

Battery-integrated wireless pressure monitoring system based on SISS

Building on the demonstrated integration capability of SISS, the battery-integrated wireless sensing system was integrated onto the outer surface of a shirt for continuous posture monitoring in long-term bedridden users (Fig. 6a). The system integrates a wireless charging coil and rectifier module, six ZIB power units, a sixteen-channel piezoresistive sensing array, and flexible PCB without the need for redundant external wiring (Fig. 6b). Figure 6c illustrates the signal-acquisition workflow of the battery-integrated wireless sensing system. The sensor voltages are collected by a micro-controller unit (MCU) equipped with five analog-to-digital converters (ADC) and five general-purpose input/output (GPIO) channels. The acquired data are then transmitted smoothly to the host device via a LoRa wireless communication module for subsequent data processing, pattern recognition, and visualization. Supplementary Fig. 33 presents the schematic diagrams of the circuit, layout design, and laser-patterning layout of the fully wireless battery-integrated back pressure monitoring system.

Fig. 6. Stretchable and breathable battery-integrated wireless pressure monitoring system based on SISS.

a Photograph of the stretchable and breathable integrated system worn on the back for user monitoring. Scale bars: 5 cm. b Photograph detailing the key functional components of the system, including the wireless charging unit, ZIBs, stretchable circuits, and piezoresistive sensors. Scale bars: 2 cm. c Schematic illustrations of the multi-channel pressure signal acquisition and analysis system including signal acquisition, processing, wireless transmission and mobile application. d Photograph showing the stable electrical performance of the integrated system under 30% biaxial tensile strain. Scale bars: 2 cm. e Visualization of pressure distribution maps and f corresponding time-sensor heatmaps under different applied pressures (2, 5, 20, 10 kPa) at 0% and 30% biaxial tensile strain. Schematic illustrations of a bedridden user, along with the corresponding pressure distribution map and time-sensor heatmap for g supine, h right-side, and i left-side postures. Photographs of the skin states after covering with j SISS- and k PPS-system for 1 h exercise. l Continuous temperature monitoring of an adult body during sleeping using SISS- and PPS-system.

Benefiting from the strain-isolation design of the SISS structure, the system maintains stable operational performance under large deformation. As shown in Fig. 6d and Supplementary Video 2, the SISS system preserves stable electrical function under a biaxial tensile strain of 30%. Furthermore, the strain stability of the sensing array is verified by pressure distribution map (Supplementary Fig. 34) and time-sensor heatmap. When pressures of 2, 5, 20, and 10 kPa are applied to channels 3, 8, 9, and 14, respectively, the pressure distribution map (Fig. 6e) and dynamic time-sensor heatmap (Fig. 6f) exhibit negligible variations between the initial state and the 30% strain condition, demonstrating excellent strain-decoupling capability and spatial distribution stability, achieving signal robustness at the system level.

The system was then employed for bedridden posture monitoring, where combining pressure distribution map and time-sensor heatmap allowed different sleeping postures to be clearly identified and comparatively analyzed (Fig. 6g–i). For example, in the supine position, the pressure is mainly distributed in the central back region and exhibits an approximately symmetric pattern (Fig. 6g). In contrast, in the right (Fig. 6h) and left (Fig. 6i) lateral positions, the pressure shift toward one shoulder back side and form a directional asymmetric pattern, accompanied by an overall increase in the pressure level compared with the supine state. A strong correspondence is observed among the photographic posture records, pressure distribution map, and time-sensor heatmap, confirming the reliable mapping between posture changes and sensing output. In addition, a time-threshold-based posture alert function was implemented to prompt users to adjust their posture when sustained localized loading is detected, thereby reduce the risk of potential tissue damage caused by prolonged localized pressure (Supplementary Fig. 35).

The wearing comfort and stability of the battery-integrated wireless sensing were also evaluated. Owing to its excellent breathability and skin compatibility, the SISS-system does not induce sweat accumulation or skin irritation during intensive exercise (Fig. 6j), with the skin moisture remaining at ~42%. In contrast, non-breathable like the PPS-system shows sweat accumulation, with the skin moisture increasing to ~70%, accompanied by noticeable erythema in the covered skin region (Fig. 6k). During continuous 6-h sleep monitoring, the SISS-system shows a thermal state close to that of natural skin situation, whereas PPS-system initially displays a significantly lower thermal level than the skin thermal environment (Fig. 6l). This behavior can be attributed to the breathable SISS allows better convective heat transfer between the air and skin while the non-breathable PPS dressing with heat conduction only. To further evaluate long-term wearability, the SISS-system was continuously worn for 3 days. As shown in Supplementary Table 4 and Supplementary Fig. 36, no obvious skin irritation is observed. The local skin humidity and temperature increase by only 7.7–9.1% and 1.28–1.53%, respectively (Supplementary Fig. 37a, b), while the SISS retains over 98% of its initial air permeability (Supplementary Fig. 37c). In addition, the integrated sensor exhibits negligible signal drift (<7.6%) throughout the test (Supplementary Fig. 38), confirming the superior long-term comfort and operational stability of the SISS-system. Overall, the SISS-system provides a breathable and stretchable platform for system-level integration of multifunctional electronics with sustained and reliable performance. Looking forward, the multilayer design strategy offers opportunities to further expand device functionality for wearable sensing and interactive applications, including physiological monitoring, electronic skins, smart textiles, and soft robotic systems.

Discussion

In conclusion, we demonstrated a breathable SISS enabled by a swelling-induced modulus design strategy for highly integrated multifunctional electronic systems with reliable mechano-electrical performance and long-term wearing comfort. Through monomer swelling and in-situ polymerization, a programmable strain-isolated island-bridge architecture was constructed while preserving the porous structure and intrinsic breathability. The obtained SISS showed high stretchability (519%, 3.86 MPa) and electrical stability (ΔR/R₀ = 3.78 under 500% strain), together with high air (31.4 mm s⁻¹) and moisture (2577.67 g m⁻² day⁻¹) permeability, surpassing previously reported stretchable and breathable supports. The SISS allowed spatially programmed integration of functional devices, with wireless charging modules, ZIBs, and piezoresistive sensing units assembled on rigid-island regions and electrical interconnections on soft-bridge areas, which remained mechano-electrical stable under 100% tensile strain. The battery-integrated wireless pressure system exhibited robust operation under 30% biaxial strains and enabled reliable posture monitoring with long-term comfortable wear. Our work provides a scalable pathway toward next-generation stretchable, breathable, and highly integrated wearable electronic systems.

Methods

Materials

All the solvents were used as received. Polyurethane (PU, Ispin), isobornyl acrylate (IBOA, Adamas-beta), butyl acrylate (BA, Adamas-beta), aliphatic urethane diacrylate (AUD, Ebecryl 8413, Allnex), phosphine oxide (TPO, Sigma–Aldrich), liquid metal (LM, eutectic GaIn, melting point, 15.7 °C; Huatai), and Ti₃C₂T_x (Jilin 11) were used as received.

Fabrication of SISS

The fabrication of the SISS involved (1) electrospinning a PU membrane and (2) selectively printing the PAp inks. Specifically, the PU electrospinning solution was prepared by dissolving PU (20 wt%) in a mixed solvent of tetrahydrofuran and dimethylformamide (1:1 v/v). Electrospinning was carried out at an applied voltage of 12 kV with a collecting distance of 15 cm. The Stiff-PAp was prepared by mixing IBOA (monomer), BA (monomer), AUD (polymerization accelerator), and TPO (photoinitiator) at a weight ratio of 5:3:2:0.1, followed by magnetic stirring for 5 h to obtain a homogeneous mixture. The Soft-PAp was prepared using the same procedure, except that the weight ratios of BA, IBOA, AUD, and TPO were adjusted to 5:3:2:0.1. Programmable patterned masks were fabricated by selectively removing the glassine paper using a power-tunable CO₂ laser and then placed onto the PU substrate for selective printing of Stiff-/Soft-PAp inks at predefined locations. The printed Stiff/Soft-PAp regions were subsequently polymerized under UV light (365 nm) for approximately 10–15 s to obtain the SISS. For comparison, PU/PDMS and PU/Ecoflex substrates were also fabricated. PDMS (Sylgard 184, 10:1) and Ecoflex (Smooth-On, 1:1) precursors were spin coated onto the PU membrane at 1000 rpm for 120 s, followed by thermal curing in an oven (80 °C for 30 min and 60 °C for 10 min, respectively).

Fabrication of functional units

The battery-integrated wireless pressure monitoring system was designed based on SISS, with a wireless charging coil as the charging module, ZIBs array as the power module, a sixteen-channel piezoresistive sensing array as the receptor, and a LoRa integration module for wireless power transmission. All electrical interconnections were realized by printing surface oxidized LM onto the soft-bridge regions using a template-assisted method without further activation. Specifically, 20 mL of LM was placed into a 50 mL glass beaker and stirred at 80 °C in air to obtain surface oxidized LM. Circuit molds were precision-cut from glassine paper using CO₂ laser and securely clamped onto a PU/Soft-PA bridges. The obtained LM was carefully printed into the mold to ensure complete filling of all corners without introducing air bubbles. After printing, the surface was then leveled by gently scraping off excess material, resulting in a well-defined circuit. The same LM template printing process was employed to fabricate the wireless charging coil onto the PU/Stiff-PA regions. The ZIB was assembled using a commercial 10 μm Zn foil as the anode, Na₂V₆O₁₆·1.63H₂O (NVO) as the cathode⁶², and a ZnSO₄/polyacrylamide-chitosan (ZnSO₄/PAM-CS) hydrogel as the electrolyte⁶³. The sensor was composed of interdigitated electrodes (Cr/Ag) and a sensing layer (PU/MXene). The Cr/Ag interdigitated electrodes were prepared via a masked vapor deposition technique onto the PU/Stiff-PA region, with a thickness of approximately 500 nm. PU/MXene was prepared by immersing a PU film in a 10 wt% MXene dispersion for 12 h, followed by natural drying at room temperature. Ti₃C₂T_X MXene was synthesized via a selective HF etching method, as described previously⁶⁴. Prior to use, the functional devices and circuit components in the battery-integrated wireless pressure monitoring system were encapsulated with PA to ensure operational stability.

Characterization

The morphologies were characterized using X-ray diffraction (XRD, Bruker D8 DISCOVER A25) and scanning electron microscopy (SEM, Zeiss, 20.0 kV) equipped with an EDX detector, respectively. Fourier transform infrared (FTIR) spectra were measured by Thermo Scientific, Nicolet iS50. A contact angle measuring device (KRUSS DSA25) was used to measure the contact angles. Raman spectra were collected using a confocal Raman microscope (WITec, Alpha300R). AFM nanomechanical characterization was performed using a Dimension Icon system (Bruker). The mechanical properties of the substrates were evaluated using a tensile machine (Instron, 3344). I–R measurements were performed with a digital source meter (Keithley 2400). The air permeability tests were performed using a fully automatic air permeability tester (YG46IF, Wenzhou Fangyuan) following the ASTM D737-08 standard (airflow pressure, 100 Pa). The moisture permeability of the substrates was measured via a TEXTEST FX3150 water vapor permeability tester in accordance with the GB/T 12704.1-2009 standard. The water vapor transmission was evaluated using the cup method under the standard ASTM E96/E96M-13. Hydrostatic pressure resistance was measured with an SDL Atlas M018 tester following ISO 811:2018. Substrate stiffness was determined using a FY207-II tester in accordance with GB/T 8942-2002. Substrate abrasion resistance was evaluated using a YG522FA tester (Wenzhou Fangyuan) equipped with 600-mesh sandpaper, following GB/T 21196.2-2007. The skin moisture was measured by a handheld skin tester (Real Bubee RBX916). Temperature characterization was performed via thermal imaging using a handheld infrared camera (FLIR C5).

Simulation and calculations

Finite element simulations were performed using Abaqus software to investigate the stress distribution of different substrates under tensile conditions. Specifically, the SISS consisted of two PU/Stiff-PA islands (7 × 7 × 0.2 mm³) connected by PU/Soft-PA bridges (18 × 7 × 0.2 mm³), distributed within a PU cuboid substrate (50 × 20 × 0.2 mm³). The experimentally determined Young’s modulus and Poisson’s ratio were assigned as the material parameters in the simulation. Subsequently, a uniaxial strain was applied along the X-axis direction to the sample. For comparison, the stress distributions of the PUS and a substrate without bridges (PUS-Stiff island) were also simulated under identical conditions. Density functional theory (DFT) calculations were performed to investigate the molecular interactions driving PU swelling in acrylate monomers. DFT calculations were conducted in Materials Studio 2024 (MS 2024) with the projector augmented-wave method. The Perdew-Burke-Ernzerhof functional described exchange-correlation effects, and van der Waals interactions were incorporated via DFT-D3 dispersion correction.

Functional characterization

Wireless charging was tested using a resonant circuit consisting of a 6.3 μH transmitting coil and a parallel 4.7 nF capacitor, driven by an alternating current (AC) signal from a RIGOL DG1032Z generator. The PU/Stiff-PA substrate with a printed LM antenna was positioned atop the coil, and the induced AC was rectified to direct current using four 1N4148 diodes. The electrochemical characterizations of ZIBs were performed using CV and GCD measurements on an electrochemical workstation (CHI 760E, Chenhua). The CV curves were recorded at different scan rates, and full-cell capacities were calculated based on the electrode area. Cycling performance was evaluated using a LAND battery testing system. The current responses of the sensors under different applied pressures were monitored using a KEITHLEY 2400 source meter.

Note: All experiments involving on-skin devices and human subjects were conducted in accordance with relevant laws and approved by the Medical and Experimental Animal Ethics Committee of Northwestern Polytechnical University (Approval No. YX202602067). This study was intended to validate the wearability and sensing performance of the battery-integrated wireless sensing system; therefore, sex (or gender) was not considered as an experimental variable. Written informed consent was obtained from all participants prior to the experiments.

Supplementary information

Supplementary Information ^{(4MB, pdf)}

Transparent Peer Review file ^{(1.8MB, pdf)}

41467_2026_76224_MOESM3_ESM.pdf ^{(68.1KB, pdf)}

Description of Additional Supplementary Files.

Download video file ^{(1.2MB, mp4)}

Supplementary Video 1.

Download video file ^{(615.7KB, mp4)}

Supplementary Video 2.

Source data

Source Data ^{(4.5MB, xlsx)}

Author contributions

C.G. supervised the whole project. J.P. and Q.C. carried out the materials synthesis and characterization. W.Z. and B.D. performed the integrated design. W.Z. and F.B. carried out calculations and simulations. J.P., Q.C., M.L., S.Z., J.C., and Y.Z. provided important experimental insights. All the authors discussed the results and contributed to writing the manuscript.

Peer review

Peer review information

Nature Communications thanks the anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Funding

C.G. discloses support for the research of this work from the National Natural Science Foundation of China (U24A20299, W2512040, 62288102), Natural Science Foundation of Shaanxi (2025RS-CXTD-003), Natural Science Foundation of Xi’an (24ZDCYJSGG0039), Natural Science Foundation of Ningbo (2024J011) and the Fundamental Research Funds for the Central Universities. Q.C. discloses support for the research of this work from the National Natural Science Foundation of China (22509158). B.F. discloses support for the research of this work from the National Natural Science Foundation of China (524B2170). J.P. discloses support for the research of this work from China Postdoctoral Science Foundation (2025M774360). W.Z. discloses support for the research of this work from Innovation Foundation for Doctor Dissertation of Northwestern Polytechnical University (No. CX2025105).

Data availability

The data that support the findings of this study are available from the corresponding author. The data generated in this study are provided in the Supplementary Information/Source Data file. Source data are provided with this paper.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41467-026-76224-w.