Strong and crack-tolerant elastomers via geometrically confined H-bonding semicarbazides

Abstract

The development of high-performance elastomers that are simultaneously strong, crack-tolerant, and wear-resistant remains a persistent challenge. Herein, we design bio-inspired semicarbazide chain extender featuring high-density hydrogen-bonding sites to synthesize poly(urethane-urea) (PUU). The use of two such extenders creates geometric confinement that promotes ordered H-bonding arrays, which synergistically enhances the mechanical performance. The resulting PUU-HI elastomer exhibits a nanoscale-ordered phase-separated structure and maximized H-bonding, achieving a tensile strength of 120.2 MPa, toughness of 400.5 MJ m⁻³ and true fracture stress of 1.3 GPa, even surpassing spider silk. The architecture additionally delivers high crack tolerance, fatigue resistance, and high wear resistance, making it ideal for stable, long-term used triboelectric nanogenerator interfaces. Solid-state NMR reveals the geometric-confinement-induced ordered and high-density H-bonding structure in hard domain for efficient energy dissipation. By designing tailored H-bonding motifs and amplifying supramolecular interactions via geometric confinement, this work offers a promising strategy for developing mechanically robust and durable elastomers.

The development of high-performance elastomers that are strong, crack-tolerant, and wear-resistant remains a challenge. Here the authors design a poly(urethane-urea) system based on geometrically confined semicarbazide chain extenders, which enhance the mechanical performance of the material.

Introduction

High-performance elastomers exhibit unprecedented application prospects in revolutionary fields, such as wearable flexible electronic devices^1,2, soft robots^3,4, and medical implants^5,6 due to their large-strain reversible stretchability. Meanwhile, the rapid development of these cutting-edge applications poses nearly stringent requirements for the high mechanical properties of materials. Especially, tensile strength represents a fundamental property in materials science, defining the ultimate load-bearing capacity of structural materials⁷. However, traditional elastomer materials (such as natural rubber, silicone rubber, etc.) often fall short in terms of performance when dealing with extreme service environments, for example, harsh mechanical wear, strong chemical corrosion, drastic temperature fluctuations, or complex dynamic loads. Therefore, the development of high-performance elastomers that combine excellent mechanical strength, outstanding toughness, superior environmental stability, and crack-tolerance has become an urgent pursuit in the field of materials science.

Consequently, polyurea-polyurethane (PUU) elastomers have emerged as promising materials with significant application potential^8–10. As block copolymers synthesized via stepwise polymerization of isocyanates, polyols, and amine chain extenders, these materials effectively integrate the flexibility of polyurethane segments with the rigidity and strong hydrogen bonding capability of polyurea segments. This distinctive microphase-separated structure imparts exceptional overall performance—exhibiting extremely high elongation at break and tensile strength^11,12 and superior abrasion and tear resistance^13,14. These attributes enable PUU elastomers to overcome the limitations of conventional elastomers, demonstrating wide-ranging applicability in advanced fields such as encapsulation layers and substrates for flexible electronic devices¹⁵, industrial rubber rollers, durable shock-absorbing and sealing components, and even biomedical implants^16,17. However, most PUU elastomers still fall short of meeting the extreme impact-resistant mechanical requirements of many next-generation technologies due to their inherent trade-off between strength and toughness.

Synthetic materials have yet to replicate the remarkable synergy of high strength and high extensibility that is so commonly realized in natural systems. Particularly intriguing is spider silk, achieves its exceptional properties through evolutionary-optimized molecular design: a hierarchical architecture with nanoscale microphase separation (~10 nm) composed of highly organized β-sheet crystalline regions crosslinked by high-density hydrogen bonds (H-bonds)¹⁸. It has a tensile strength of up to 1.5 GPa, a toughness of ~350 MJ m⁻³, and at the same time maintains environmental degradability^19,20. Inspired by this, H-bonding network engineering has become the core strategy to break through the performance bottleneck of polyurethane elastomers^10,21–25. For example, H-bonded 2ureido-4[1H]-pyrimidinone (UPy) dimer has been widely used to construct strong and tough elastomers^26–31. In particular, an increasing number of studies have reported that hydrazide, serving as a chain extender instead of traditional diamine for PUU and providing additional H-bonding sites, can significantly enhance mechanical properties. For example, Sun et al. prepared PUU with a higher H-bond density using hydrazide-terminated monomers with stronger H-bonding ability³². Wang et al. developed a general design principle for toughening polymer materials by mismatched supramolecular interactions, propelling elastomers into the GPa-class mechanical regime³³. These pioneering works fully demonstrate the great potential of H-bond engineering in improving the performance of PUUs. Nevertheless, achieving extreme enhancement of H-bond density coupled with precise structural control over phase separation to realize super-strong elastomers remains a central, yet unresolved challenge in this field.

In this work, drawing inspiration from the high-density H-bonding assembly in spider silk, we present a design strategy employing semicarbazide-based chain extenders. Furthermore, we introduce a geometric confinement principle through the combined use of two such extenders (N, N'-(methylenebis(cyclohexane-4,1-diyl)) bis(hydrazine-carboxamide and N-((3-(hydrazinecarboxamido)-1-methylcyclohexyl)methyl) hydrazinecarboxamide, abbreviated as HMDAA and IPDAA, respectively), aiming to amplify supramolecular interactions, to yield a robust PUU-HI elastomer with high crack tolerance (as illustrated in Fig. 1). Such hybrid semicarbazide-based chain extenders with optimal ratio could spontaneously interlock and enable programmable dense H-bond networks through geometrically confined assembly. This design can force highly ordered close-packed hard-segment assemblies, simultaneously maximizing H-bond density. Consequently, the resulting PUU-HI elastomer exhibits exceptional mechanical performance, including a tensile strength of 120.2 MPa, toughness of up to 400.5 MJ m⁻³ (with a true fracture stress of 1.3 GPa), surpassing the overall mechanical properties of high-quality natural spider silk. Molecular dynamics simulation revealed the geometric confinement-induced enhancement of H-bonding interactions in PUU-HI. Multidimensional solid-state NMR reveals the formation of ordered H-bonds, microphase separation, and sterically restricted segmental mobility due to geometric confinement of two semicarbazide-based chain extenders. These NMR findings provide an in-depth understanding of the microscopic origin of high mechanical properties in PUU elastomers at the atomic level. Due to the reversible and dynamic nature of the densely packed H-bonded hard segments, the material also demonstrates excellent tear resistance, resilience, and recyclability under mild conditions. This strategy resolves the strength–toughness trade-off and enables durable, wear-resistant elastomers for repeated-contact applications, including triboelectric nanogenerators (TENG), offering a universal molecular design framework.

Fig. 1. Bio-inspired design and synthesis of PUU elastomers with semicarbazide chain extenders.

Schematic view of a the multi-level structure of spider silk and the high-density hydrogen-bond aggregates in it, and b three generations of amino-based chain extenders for PUU synthesis. c Molecular structure of semicarbazides IPDAA and HMDAA (denoted as “I” and “H”) and synthetic routes of the PUU-H, PUU-I, and PUU-HI elastomers. d Schematic illustration of the hydrogen-bonding network in PUU-HI.

Results

Preparation of PUU-HI elastomers

Two semicarbazide derivatives (termed HMDAA and IPDAA) were synthesized from diisocyanates (Supplementary Figs. 1–6), used as the chain extenders. Several different types of linear PUU polymers were designed and synthesized from the hydroxyl-terminated polytetramethylene ether glycol (PTMEG), diisocyanates, and a semicarbazide (or mixture of two bis-semicarbazides) chain extender via a two-step reaction process (Fig. 1c). The obtained PUU samples, designated as PUU-H_xI_y, where x and y represent the molar ratios of HMDAA and IPDAA (Supplementary Table 1 for corresponding feed ratios), were used for subsequent experiments. To investigate the effect of the hard segment on the properties of PUU elastomers, control samples with the aliphatic isocyanate hexamethylene diisocyanate (HDI) and the alicyclic isocyanate 4,4′-dicyclohexylmethane diisocyanate (HMDI) were synthesized, yielding PUUHI-HDI and PUUHI-HMDI. The chemical structures of the PUU polymers were confirmed by the ¹H NMR analysis and Fourier-transform infrared (FTIR) spectroscopy (Supplementary Figs. 8–11). The disappearance of the absorption peak at 2264 cm⁻¹ confirms the complete reaction of the -NCO functional group, while the appearance of the -N-H peak at 3340 cm⁻¹ proves the formation of the urea group. In addition, the signals between 1750 and 1600 cm⁻¹ are attributed to the carbonyl group, which will be used to analyze the formation of H-bonds in PUU in the following text. To compare the mechanical properties of the samples synthesized with amine, hydrazine, and semicarbazide chain extenders, a set of control experiments was set up. Such semicarbazide derivatives can significantly increase numbers of H-bonding sites in the obtained PUU polymers (PUU-HDAA, Supplementary Fig. 12), compared to those synthesized from the commonly used diamine (PUU-HMD) and previously used hydrazide chain extenders (PUU-SDH). The molecular weight and distribution indices were evaluated by GPC (Supplementary Fig. 13 and Table 2), all samples have similar molecular weights and molecular weight distributions.

Mechanical properties of the PUU elastomers

Mechanical properties of the PUU elastomers with varying chain extender ratios were characterized through tensile tests at 50 mm min⁻¹. The obtained key mechanical performance parameters are summarized in Supplementary Table 3. PUU-HMD (with the hard segment structure shown in Fig. 2a), chain-extended with diamine, exhibits a fracture strength of 55.3 ± 4.9 MPa, a fracture strain of 1203.6 ± 147.9%, and a toughness area of 205.6 ± 13.9 MJ m⁻³ (Fig. 2b and Supplementary Fig. 14). In comparison, the elastomer PUU-SDH, chain-extended with hydrazide, demonstrates a fracture strength of 76.9 ± 2.4 MPa, a fracture strain of 1104.8 ± 31.2%, and a toughness area of 233.3 ± 23.9 MJ m⁻³, which is 1.1 times that of PUU-HMD. Notably, PUU-HDAA chain-extended with semicarbazide achieves a remarkably high fracture strength of 87.1 ± 3.5 MPa, a fracture strain of 1062.4 ± 19.4%, and a toughness area of 317.3 ± 18.7 MJ m⁻³, representing 1.6 times that of the PUU-HMD elastomer. The semicarbazide-based chain extenders, benefiting from an increased number of H-bonding sites, facilitate the simultaneous enhancement of both strength and toughness by mimicking the dense H-bonding network characteristic of spider silk. Compared with the previously reported alcohol, amine, and acyl hydrazine chain extenders, it can be found that this type of chain extender has significant advantages in constructing strong PUU elastomers (Fig. 2c)^16,32–40.

Fig. 2. Mechanical properties of PUU elastomers.

The hard segment structures (a) and typical stress–strain curves (b) of PUU-HMD, PUU-SDH, and PUU-HDAA elastomers, measured at the strain rate of 50 mm min⁻¹. c Tensile strength as a function of the H-bonding forms of PUU-HDAA in comparison with the results of the PUs reported in the literature. d The stress–strain curves of PUU-HI elastomers with different ratios of H/I chain extenders. e Comparison of the tensile strength and toughness among the PUU-HI and some of the literature-reported elastomers. f True stress–strain curves of PUU elastomers. g Comparison of the true stress at break and toughness among the PUU-HI elastomer and different types of spider silk. h The corresponding tensile strength and toughness of PUU elastomers with different diisocyanates. i Photograph of a PUU-HI elastomer film showing high transparency.

PUU-I and PUU-H have a tensile strength (i.e., ultimate engineering stress) of 72.6 ± 7.4 MPa and 79.3 ± 0.5 MPa, an elongation-at-break of 1149.3 ± 33.4% and 780.9 ± 23.6%, resulting in a toughness of 229.6 ± 8.5 MJ m⁻³ and 244.2 ± 6.8 MJ m⁻³, respectively (Fig. 2d and Supplementary Fig. 15). It is noted that equimolar incorporation of geometrically confined HMDAA and IPDAA chain extenders produced PUU-HI with strong mechanical properties: a high tensile strength of 120.2 ± 1.1 MPa and toughness of 400.2 ± 20.7 MJ m⁻³. To our knowledge, the tensile strength of PUU-HI exceeds that of both commercial rubbers and polymer elastomers reported in the literature, while maintaining a high level of toughness (Fig. 2e)^{32,33,35,38,41–48}. This indicates that HMDAA and IPDAA reach the optimal synergistic matching state at a ratio of 1:1. When the molar ratio of HMDAA:IPDAA deviates from 1:1, both the toughness and strength decrease to varying degrees, but there is still a significant improvement compared to using a single chain extender. This result can be explained by the good compatibility between hard segments formed by two semicarbazides. An obvious strain hardening phenomenon⁴⁹, associated with the maximum tensile strength, can be observed following the distinct upturn (at 600–700% strain, Supplementary Fig. 16) in the slope of the stress–strain curves. At this stage, the sample exhibits increased stiffness, along with a significant enhancement in modulus and strength. This behavior can be attributed to the progressive rearrangement and alignment of H-bonded hard segments under tensile deformation. Notably, PUU-HI demonstrates the highest slope during the strain hardening region, which is likely due to its higher density of H-bonds compared to other samples. Similar mechanical behavior has been reported in spider dragline silk, where the unfolding of β-sheet nano-crystals is responsible for the strain-hardening response⁵⁰. In addition, tensile tests at varying strain rates (Supplementary Fig. 17) revealed that PUU-HI exhibits stable and robust mechanical performance over a wide range of strain rates. The observed increase in tensile strength with increasing strain rate and the decrease in elongation at break reflect the dynamic nature of the hydrogen-bonding network, demonstrating the efficient energy dissipation and structural stability imparted by the densely packed hard domains.

The stress–strain curve provides a more accurate indication of the capacity to withstand external stress for highly ductile materials with large deformability. The true stress–strain curve (Fig. 2f) indicates that the true fracture stress of PUU-HI reaches 1.3 GPa, which is notably higher than that of PUU-I and PUU-H elastomers. In terms of both true fracture stress and toughness, the PUU-HI elastomer surpasses various types of spider silk (fracture stress: 0.8–1.5 GPa; toughness: 160 MJ m⁻³). Specifically, its true fracture stress exceeds that of most spider silks (Fig. 2g), and its toughness (400.5 MJ m⁻³) is approximately 2.5 times greater than that of typical spider silk, even outperforming the toughest known natural fiber, Darwin’s bark spider silk (~354 MJ  m⁻³)⁵¹. Therefore, the PUU-HI elastomer exhibits superior mechanical robustness, including both true stress at break and toughness compared to typical spider silk. Furthermore, the influence of different diisocyanates on the mechanical properties of PUU was investigated. Among them, PUU-HI with IPDI exhibits the highest tensile strength and toughness (Fig. 2h and Supplementary Fig. 18). This sample also demonstrates colorlessness and good transparency (Fig. 2i), with a transmittance exceeding 85% in the visible light spectrum (Supplementary Fig. 19). This optical transparency, along with its high mechanical properties, makes PUU-HI a promising candidate for flexible transparent electronics, including transparent triboelectric nanogenerators and other durable optoelectronic devices.

High-density H-bonds induced by geometric confinement

In order to explore the H-bonding interactions in elastomers, the contributions of urethane groups and urea groups in the elastomer structure to the formation of H-bonds in thermoplastic elastomers were studied in detail by FTIR spectroscopy. A detailed peak fitting analysis was carried out for the region of 1750–1600 cm⁻¹ (Fig. 3a–c and Supplementary Fig. 20). Typically, the broad absorption bands in the range of 1750–1550 cm⁻¹ can be assigned to the C=O stretching region of urethane(-NH-COO-) and urea(-NH-COO-NH-). These absorption bands can deconvoluted into several subpeaks, including the free C=O (I) and the H-bonded C=O (II) in the urethane groups, and the free C=O (III) and H-bonded C=O (VI, V, VII) in the urea groups^31,52. Quantitative analysis of the peak fitting (Supplementary Table 4) revealed that the content of H-bonded C=O follows the order: PUU-HDAA (62.3%) > PUU-SDH (48.7%) > PUU-HMD (38.2%). This indicates that semicarbazides, as next-generation chain extender can significantly enhance the mechanical properties of PUU due to its ability to form high-density H-bonds. In addition, the content order of the H-bonded C=O in samples with various semicarbazide chain extenders is PUU-HI (71.4%) > PUU-H (64.6%) > PUU-I (46.7%) (Supplementary Table 5). In fact, the content of H-bond donors and acceptors in the three samples is identical, and the theoretical capacity for H-bond formation is essentially the same, yet PUU-HI exhibits the higher H-bond content. This result suggests that the hybrid chain extenders in PUU-HI enable a geometrically closer molecular packing, which in turn facilitates the formation of H-bond arrays involving denser H-bonds. This structural feature also explains the enhanced tensile strength and energy dissipation performance of the PUU-HI elastomer (Fig. 2c).

Fig. 3. Structural characteristics PUU elastomers.

a–c FTIR spectra in the C=O stretching vibration region and d 1D SAXS profiles of PUU-H, PUU-I, and PUU-HI elastomers. e AFM pattern of PUU-HI. f DSC curves of all PUU samples. g, h Snapshots of the all-atom MD simulations of the structures of PUU-HI and PUU-I. i The counted number of H-bonds in PUU-b, PUU-H or I, and PUU-HI.

Due to the thermodynamic incompatibility between the soft and hard segments, polyurethane elastomers usually spontaneously form a microphase separation structure, which could be facilitated by strong H-bonding interactions. Small-angle X-ray scattering (SAXS) was employed to further study the phase separation structure of PUU elastomers. As can be seen from the SAXS patterns, PUU-I and PUU-H have relatively broad and weak peaks at q ~ 0.8 nm⁻¹ and q ~ 0.58 nm⁻¹, respectively, and the corresponding long period sizes are 7.8 and 10.8 nm. PUU-HI exhibits a relatively narrow and well-defined scattering peak at q ~ 0.75 nm⁻¹, corresponding to a periodic length around 8.5 nm. This indicates that PUU-HI possesses a higher degree of phase separation and improved local ordering. The hard domains, characterized by a high density of H-bonds, act as a reinforcing agent and provide physical crosslinking points at the nanoscale, which significantly contribute to the material’s mechanical performance. Notably, the phase domain size of PUU-HI is comparable to that of spider silk⁵⁰, which is considered a key microstructural factor underlying its high mechanical properties. The atomic force microscopy (AFM) image of the PUU-HI sample further illustrating its phase-separated morphology (Fig. 3e). Differential scanning calorimetry (DSC) curves of five PUU samples (Fig. 3f) shows that PUU-HI exhibits the highest melting temperature, which is also attributed to the higher density of H-bonds, regular and orderly phase separation structure. Beyond tensile properties, this densely H-bonded microstructure also confers inherent resistance to surface damage under repeated contact. The combination of high H-bond density and ordered hard domains enables efficient stress distribution and energy dissipation at the interface, which are essential for wear-resistant applications.

All-atom molecular dynamics (MD) simulations were employed to theoretically investigate the influence of hybrid H-bonding motifs on the microstructure of PUU elastomers. Representative snapshots from the all-atom MD simulations are displayed for PUU-HI and PUU-I (Fig. 3g, h). By analyzing the number of H-bonds in the simulated structures as shown in Fig. 3i, it was observed that PUU-HI contains a higher number (~142) compared to PUU-I (~133) and PUU-H (~128) (Supplementary Fig. 22a), which aligns with the results obtained from infrared peak fitting. Moreover, these values are significantly higher than that of the control sample PUU-b without semicarbazide chain extender (~56, see Supplementary Figs. 21 and 22b). This indicates that semicarbazide-based chain extenders can substantially enhance the H-bond density within the polymer matrix. Moreover, FTIR and MD finding provides evidence that the strategy of using hybrid chain extenders directly enhances the supramolecular interactions in PUU elastomers. As a result, PUU-HI exhibits a higher degree of H-bonding packed hard domains, which contribute to the enhanced strain hardening behavior and superior mechanical performance of PUU-HI.

Multidimensional solid-state NMR experiments were further used to clarify the atomic-level mechanisms of geometric complementarity and H-bonding, and heterogeneous dynamics in PUU samples. Figure 4a presents a schematic representation of the molecular architecture of the hard segment region and probable H-bonds, along with the assignment of selected ¹³C nuclei. ¹³C cross polarization/magic angle spinning (CP/MAS) NMR (Supplementary Fig. 23a) spectra revealed structural distinctions among the samples (Fig. 4b). Notably, the spectral features of the carbonyl carbon resonance within the 150–170 ppm range (Fig. 4c) are highly sensitive to H-bonding interactions⁵³, enabling further analysis of such interactions. PUU-cast and PUU-blend refer to samples prepared by casting and blending PUU-H and PUU-I in DMF solution and through physical blending, respectively. The carbonyl carbon signals of PUU-HI and PUU-cast nearly overlap, whereas a significant shift is observed in the PUU-blend sample (Fig. 4c). This observation confirms the geometric complementarity induced enhancement of H-bonding interactions between the two semicarbazide-based chain extenders. The ¹³C chemical shift anisotropy (CSA) provides precise insights into extranuclear electron density variations and serves as a sensitive probe for analyzing molecular structures and dynamics⁵⁴. To detect the difference of H-bonding structure in the hard domain for PUU samples, the ¹³C CSA powder spectrum of the H-bonded C=O was obtained through a two-dimensional (2D) ¹³C SUPER experiment (Supplementary Fig. 23b), in which the F1 dimension represents the refocused CSA powder pattern of each ¹³C site, and the F2 dimension is the high-resolution ¹³C spectrum. By fitting the CSA curves (Supplementary Eqs. 1–6) extracted at the ~160 ppm carbonyl carbon signal from the 2D ¹³C SUPER spectra (Fig. 4d) of PUU-HI and PUU-blend samples (Fig. 4e), the CSA parameter (δ_CSA) and the asymmetry parameter (η) were obtained. Notably, PUU-HI exhibits obvious smaller δ_CSA and η values than that of PUU-blend, suggesting a geometric confinement effect between semicarbazide-based chain extenders, which facilitates the formation of more ordered and densely packed H-bonded hard domains.

Fig. 4. Multidimensional solid-state NMR research of PUU-HI elastomers.

a The molecular structure diagram of the hard segment region and the assignment of typical ¹³C nucleus. ¹³C CP/MAS full spectra (b) and enlarged ¹³C CPMAS NMR spectrum of carbonyl groups between 150 and 170 ppm (c) of PUU-HI, PUU-blend, and PUU-cast samples. d 2D ¹³C SUPER spectrum (contour plot) at ¹³C of 45.4 ppm. e Slice of 2D ¹³C SUPER spectrum at ¹³C of 160 ppm for PUU-HI and PUU-blend. f 2D ¹³C−¹H WISE spectrum of PUU-HI. Slices of the 2D WISE spectrum at (g) different carbon sites within PUU-HI, in which the corresponding ¹³C chemical shifts (ppm) are indicated in the figure legend, h C2 of 45.4 ppm and i C4 of 32.5 ppm for PUU-H, PUU-I, and PUU-HI elastomers. All solid-state NMR experiments were performed at 25 °C on a 400 MHz spectrometer.

Such effects could couple and mutually influence the dynamics of two distinct components (HMDAA and IPDAA), which can be further elucidated by the 2D ¹³C–¹H wide-line separation (WISE) NMR method (Supplementary Fig. 23c). Components with different mobility are distinguished by their ¹H wide-line spectra (F1 dimension): broad lines indicate rigid segments, while narrow lines correspond to motionally averaged segments⁵⁴. The associated ¹³C chemical shifts (F2 dimension) reveal the composition of corresponding segments. The WISE spectrum of PUU-HI (Fig. 4f) and the corresponding F1 dimension slice projections at specific carbon sites (Fig. 4g) reveal distinct segmental mobility. The peak at C1 (71.2 ppm), assigned to the—CH₂-O- signal from the PTMEG, the line width (ν_1/2) is quite narrow (~5 kHz), indicating soft segment is quite mobile. The peak at C2 (45.4 ppm) is attributed to the ¹³C resonance of -CH₂-NH-, the linewidth of which is quite broad (ν_1/2 ~ 54–57 kHz), demonstrating that the hard segments are quite rigid due to the enhanced H-bonding interactions in PUU-HI. The signal at C3 (36.2 ppm) is attributed to the carbon atoms in the intermediate linkage of the hard segment HMDAA as well as the -CH₂- groups in IPDI. Lastly, the peak at C4 (32.5 ppm) represents the -CH₂- groups on the six-membered ring of HMDAA and the two tertiary carbons present in IPDAA. The slices at the C2 position of various PUU samples (Fig. 4h), which are closest to the H-bond arrays within the hard domains, show that PUU-HI exhibits the largest linewidth. This reflects the strongest intermolecular interactions among the hard segments, arising from the strongest dipolar coupling of protons in PUU-HI. This suggests that geometrically confined chain extenders facilitate the formation of a denser H-bonds and stronger intermolecular interactions, thereby enhancing the mechanical properties. Furthermore, the slice spectrum of the C4 position in the aliphatic chain of PUU-I displayed a relatively narrow Lorentzian line shape (ν_1/2 ≈ 35.5 kHz). That is to say, in the absence of spatial hindrance caused by geometrically confined motifs, the IPDAA units retain relatively high molecular mobility even within the hard segment region. It was noteworthy that when the geometric confined HMDAA is introduced into PUU-HI, the linewidth of the C4 slice (Fig. 4i) becomes a completely rigid spectrum (ν_1/2 ≈ 50.6 kHz), and the characteristic sharp Lorentzian peak of C4 in PUU-I disappeared. The HMDAA molecules in PUU-H display complete rigidity (ν_1/2 ≈ 50.8 kHz), which contrasts with the faster molecular motion of the IPDI groups in PUU-I. These NMR results unambiguously demonstrate that the rigidity of hard segments in PUU-HI results from intermolecular interactions between geometrically interlocked two chain extenders with different mobility. This effect results in increased steric hindrance between the two motifs, which restricts the local motion of the mobile components and imparting a rigid structural characteristic.

The thermal stability of the samples was evaluated by DSC, thermogravimetric analysis (TGA), and dynamic mechanical analysis (DMA). DSC revealed a soft-segment glass transition (T_g) below −70 °C (Supplementary Fig. 24), indicating high chain mobility. Moreover, the confined H-bonding extenders significantly enhanced thermal stability, as evidenced by a higher temperature at 5% mass loss in TGA (Supplementary Fig. 25) and a correspondingly elevated hard-segment T_g in DMA (Supplementary Fig. 26). This confirms that geometric confinement reinforces intermolecular interactions, thereby improving thermal stability. These results collectively demonstrate that the strategy leveraging the semicarbazide chain extenders under a geometric confinement principle successfully creates a densely H-bonded and highly ordered microstructure, which is the fundamental origin of the strong and tough mechanical properties.

Elastic restorability of the PUU-HI elastomers

Owing to the densely arranged H-bonded hard domains that serve as robust physical crosslinks, PUU-HI exhibits not only high tensile strength but also substantial energy dissipation capacity. The energy dissipation behavior of the PUU elastomers was evaluated via cyclic tensile tests under increasing strain (Fig. 5a and Supplementary Fig. 27). The dissipated energy, defined as the hysteresis area of the loading-unloading curves^55,56, increased with applied strain (Fig. 5b). In the low-strain regime (<300%), all samples displayed comparable energy dissipation, which is attributed to the elongation of soft segment polymer chains and the reversible dissociation of weak H-bonds³³. However, when the strain exceeded 300%, the dissipated energy of PUU-HI showed a significant non-linear increase, and the increase amplitude far exceeded that of the control group samples. This phenomenon clearly indicates that the gradual rupture of strongly ordered H-bonded hard domains with geometric complementarity under large deformations provides higher energy absorption, fully verifying the design concept of its reinforcement and toughening. Cyclic tensile curves of the PUU-HI sample at 100% strain, measured with different waiting times between consecutive loading-unloading cycles, reveal a pronounced hysteresis loop in the first cycle (Fig. 5c). This hysteresis, corresponding to an energy dissipation of 3.46 MJ m⁻³, indicates the material’s notable energy dissipation capability. Following a resting period of 30 min, the mechanical curve recovers to 75.1% of its initial state; after 240 min of rest, the recovery increases further to 90.8%. In comparison, PUU-H, which contains only a single chain extender, achieves a recovery of merely 85% after 240 min of rest (Supplementary Fig. 28). This suggests that the robust H-bonded physical crosslinking domains contribute to the superior elastic recovery performance of PUU-HI while preserving high energy dissipation capacity.

Fig. 5. Elastic restorability of PUU-HI elastomers.

a The cyclic loading-unloading tensile curves of PUU-HI elastomer with increased strain. b The energy dissipation of PUU elastomers varying with strain during the cyclic tensile process. c Cyclic loading-unloading tensile curves of PUU-HI elastomer with different waiting times. d 2D SAXS patterns and f 1D SAXS profiles of the PUU-HI elastomers at different strains and the sample released to 0%. e Schematic diagram of chain segment evolution within PUU-HI elastomer during the stretching process. g Photographs of the initial size of the PUU-HI sample and its recovery after stretching.

Two-dimensional (2D) SAXS patterns and their corresponding one-dimensional (1D) intensity distribution curves, as well as Azimuthal intensity distribution (Supplementary Fig. 29) were collected from the PUU-HI sample upon stretching to various strains and subsequent release (Fig. 5d, f). The pristine specimen exhibited isotropic diffraction rings in the 2D scattering pattern, indicating uniform structural periodicity. Upon applying uniaxial strain to the PUU-HI elastomer, the diffraction pattern transitioned from circular to elliptical symmetry. Concurrently, the long-period spacing initially increased while diffraction peak intensities progressively diminished. At strains exceeding 500%, complete disappearance of scattering rings confirmed disruption of the ordered microstructure. This phenomenon aligns with observations by Miwa et al., where dissociation of hard-segment aggregates liberates reconfigurable H-bonding sites, thereby mitigating localized stress concentration within the polymer matrix⁵⁷.

Remarkably, the original 2D diffraction pattern and 1D intensity profile (Fig. 5f) were fully restored upon stress release and rest for 6 h. This structural recovery is ascribed to the reformation of H-bonded elastic percolating networks through dynamic H-bond dissociation/recombination processes following load removal. These observations indicate that the densely packed hydrogen-bonded hard domains in PUU-HI are interconnected to form an elastic percolating network, which enables efficient stress transfer and drives macroscopic recovery after large deformation. A mechanism for the elastic restoration of PUU-HI was proposed (Fig. 5e): (i) geometrically confined H-bonds organize highly ordered hard domains that interconnect into an elastic percolating network in the pristine state; (ii) soft segment chains align and the long period increases at small deformations; (iii) under large strain, hydrogen bonds and ordered structures partially dissociate, while upon stress removal, the percolating network facilitates rapid reformation of hydrogen bonds and recovery of microphase separation. Macroscopically, the PUU-HI elastomer fully recovered its original dimensions after being stretched to 600% strain and resting for 4 h (Fig. 5g), validating this recovery mechanism.

Fatigue resistance and crack tolerance of PUU-HI elastomers

To ensure the long-term reliability of elastomers in practical applications, good fatigue resistance and crack tolerance are crucial. To evaluate the fatigue resistance of the PUU-HI elastomer, cyclic tensile tests were further conducted under 300% strain (Fig. 6a). After the second cycle, a 26% residual strain was observed. Although some energy dissipation and plastic deformation occurred initially, the hysteretic energy dissipation quickly decreased and stabilized as the number of cycles increased. Cyclic tensile tests (1000 cycles) were conducted on PUU-HI at a constant strain of 100% and a crosshead speed of 50 mm min⁻¹ (Fig. 6b, c). A significant decrease in the maximum force of PUU-HI was observed at the onset of stretching, which we attribute to partial hydrogen-bond rupture. In subsequent stretching cycles, the maximum force plateaued, indicating that a dynamic equilibrium between hydrogen-bond dissociation and rapid re-association was established within the elastomer. This intrinsic self-recovery significantly mitigates the accumulation of damage, thereby conferring cyclic stability and fatigue resistance upon PUU-HI.

Fig. 6. Fatigue resistance and crack tolerance of PUU-HI elastomers.

a The 20-cycle tensile curves of PUU-HI elastomer at a strain of 300%. b The 1000-cycle tensile curves of the PUU-HI elastomer at a strain of 100%. c Variation of the force of the PUU-HI elastomer held at 100% strain for 1000 cycles. d Photographs of the notched PUU-HI elongated to different strains. e Notched tensile test curves of PUU-HI elastomer. f Notched tensile test curves of PUU-HI elastomer with notches to lift a 10,000-time weight of its own weight. g Force-displacement curve of PUU-HI measured by trouser tear tests. Inset: illustration of the PUU-HI stripe in the trouser tear test. h Comparison of the tear energy of PUU-HI with those of elastomers reported in the literature.

Tear resistance is a key safety performance indicator for elastomers, especially when employed in complex structures or defective environments prone to stress concentration. To evaluate the tolerance of PUU-HI elastomer to defects and damage, a notched tensile test was performed. Pre-damaged PUU-HI specimens exhibited significant crack blunting during tensile deformation (Fig. 6d). Under large-strain conditions, cracks primarily underwent transverse opening (width expansion) with minimal lateral propagation. Notably, crack extension perpendicular to the loading direction was effectively suppressed. Specimens maintained structural integrity until achieving an ultimate strain of 640%, while retaining a fracture strength of 23 MPa (Fig. 6e). Calculated via the Greensmith method⁵⁸, the fracture energy reached 303.6 kJ m⁻². This high fracture energy is attributed to the strong constrained supramolecular binding forces and energy dissipation capacity. The notched PUU-HI elastomer film is still capable of lifting a weight 10,000 times its own even in the presence of cracks (Fig. 6f). Further trouser tear tests showed that the tear energy of the PUU-HI sample was 392.1 kJ m⁻², which was significantly higher than that of most elastomeric materials reported in the literature (Fig. 6h)^{32,35,40,42,43,47,59–62}. This indicates that this elastomer has mechanical robustness and high crack tolerance. The good properties of the PUU-HI elastomer can be attributed to the dynamic fracture and rapid recombination of high-density H-bonds in the ordered hard domains. Therefore, the stress concentrated at the crack tip can be effectively transferred to the entire polymer network, thereby suppressing the lateral growth of the crack.

Wear-resistant PUU-HI for durable triboelectric nanogenerator interfaces

The enhanced crack tolerance and fatigue resistance of PUU-HI (Fig. 6) arise from its geometrically confined, high-density H-bonding network, which efficiently dissipates energy and suppresses crack propagation under mechanical stress. These same attributes are critical for triboelectric layers, which must withstand repeated frictional contact without surface degradation. We therefore explored the PUU-HI elastomer as a triboelectric layer in a TENG. Under identical testing conditions, PUU-HI exhibits a more stable coefficient of friction (µ) compared to PUU-HMD, with minimal variation during repeated rubbing (Fig. 7a). This stability reflects the superior wear resistance of PUU-HI, indicating that its surface roughness remains largely unchanged during cyclic contact, thereby helping to maintain stable interfacial contact and ensuring consistent electrical output of the TENG over long-term operation.

Fig. 7. Working mechanism and stable electrical output performance of PUU-HI TENG.

a Friction coefficient as a function of sliding time for PUU-HI and PUU-HMD under reciprocating sliding conditions. b Schematic of the operational mechanism of PUU-HI TENG with layered structure. c Output V_oc and d Q_sc of the PUUHI-TENG after 500 cycles of alternative contact/separation movements.

A flexible TENG (denoted as PUU-HI TENG) was fabricated using PUU-HI as the positive triboelectric layer, polytetrafluoroethylene (PTFE) as the negative triboelectric layer, and copper tape as the electrodes (as illustrated in Fig. 7b). The electrical output performance of the PUU-HI TENG was evaluated at contact frequencies ranging from 0.25 to 2.5 Hz. The open-circuit voltage (V_oc), short-circuit current (I_sc), and transferred charge (Q_sc) reached maximum values of −55 V, 1.1 μA, and 20 nC, respectively (Supplementary Fig. 30). To evaluate whether the material’s wear resistance translates into stable interfacial contact under repeated operation, the device was subjected to 500 continuous contact–separation cycles at 2.0 Hz. Notably, the PUU-HI TENG maintained highly stable electrical output throughout (Fig. 7c, d), demonstrating excellent operational reliability. This consistent performance, with negligible decay over 500 cycles, directly reflects the superior wear resistance and interfacial stability of PUU-HI—its surface remains intact and its friction coefficient stable (Fig. 7a), ensuring reliable charge transfer over long-term use. As a proof of concept, the PUU-HI TENG was attached to the sole of a tester’s shoe, where it successfully detected different motion states through variations in the V_oc signal (Supplementary Fig. 31). These results highlight the potential of the PUU-HI elastomer as a robust and durable triboelectric material for energy harvesting and self-powered sensing.

Discussion

In summary, drawing inspiration from spider silk, this work introduces a geometric confinement strategy that utilizes semicarbazide-based chain extenders to create strong, tough elastomers. Atomic-level multidimensional solid-state NMR and MD simulations reveal that geometric confinement forces highly ordered hard-segment stacking, which concurrently enhances H-bond density and dynamic energy dissipation. This densely H-bonded, highly ordered structural arrangement imparts high crack tolerance to PUU-HI, as demonstrated by a high tear energy of 392.1 kJ m⁻², which represents the highest value reported for elastomers to date. Beyond crack tolerance, the high-density H-bond network and geometrically confined hard domains endow PUU-HI with good wear resistance and interfacial stability under repeated contact, key attributes for demanding, friction-intensive applications. The reversible H-bonded assemblies (Supplementary Figs. 32–36) also contribute to fatigue resistance, self-recovery, and enable recyclability with >90% property retention (Supplementary Figs. 37–41). This work presents a strategy for creating strong, tough elastomers through the combined use of geometrically confined semicarbazide‑based chain extenders and tailored H‑bonding arrays. This concept of harnessing geometric confinement to amplify supramolecular interactions is expected to guide the design of durable supramolecular materials for demanding engineering applications.

Methods

Materials

Polytetramethylene ether glycol (PTMEG, M_n = 2000 g mol⁻¹), isophorone diisocyanate (IPDI, 99%), methylene-bis(4-cyclohexylisocyanate) (HMDI, 99%), hexamethylene diisocyanate (HDI, 98%), and dibutyltin dilaurate (DBTDL, 95%) were purchased from Sigma-Aldrich (USA). NH₂NH₂·H₂O (80% in H₂O) was obtained from Aladdin. Anhydrous N, N-dimethylformamide (DMF) was obtained from Shanghai Energy Chemical Co. All materials were used as received without further purification.

Synthesis of N-((3-(hydrazinecarboxamido)-1-methylcyclohexyl)methyl) hydrazinecarboxamide (IPDAA)

Hydrazine hydrate (6 g) was charged into a three-necked flask followed by the addition of isopropanol (23.4 g). The resulting mixture was stirred at room temperature for 30 min. Subsequently, a solution of isophorone diisocyanate (IPDI, 5 g) in tetrahydrofuran (THF, 43.5 g) was added dropwise slowly to the reaction mixture. Stirring was continued for an additional 3 h. The whole reaction was conducted under N₂ atmosphere. The reaction mixture was evaporated under vacuum to afford a white solid in 90% yield. ¹H NMR (400 MHz, DMSO-d6) δ 6.85 (d, J = 9.3 Hz, 2H), 6.25 (d, J = 7.7 Hz, 1H), 5.96 (d, J = 8.4 Hz, 1H), 4.06 (d, J = 24.5 Hz, 4H), 1.62–1.36 (m, 2H), 1.15–0.73 (m, 16H). ¹³C NMR (101 MHz, DMSO) δ 160.8, 159.92, 62.48, 53.21, 47.35 (d, J = 9.4 Hz), 46.91 (d, J = 11.7 Hz), 36.67, 35.49, 31.96, 28.06, 25.96, 23.76.

Synthesis of N, N′-(methylenebis(cyclohexane-4,1-diyl)) bis(hydrazine-carboxamide) (HMDAA)

Hydrazine hydrate (NH₂NH₂·H₂O, 6 g) was charged into a three-necked flask, followed by the addition of isopropanol (23.4 g). The resulting mixture was stirred at room temperature for 30 min. Subsequently, a solution of hexamethylene diisocyanate (HMDI, 5 g) in tetrahydrofuran (THF, 43.5 g) was added dropwise slowly to the reaction mixture. Stirring was continued for an additional 3 h. The whole reaction was conducted under N₂ atmosphere. The reaction mixture was then filtered, and the collected solid was washed with isopropanol to afford a white solid in 93% yield. ¹H NMR (400 MHz, DMSO-d6) δ 6.83 (d, J = 7.9 Hz, 2H), 6.22 (d J = 8.0 Hz, 2H), 4.05 (d, J = 24.5 Hz, 4H), 3.67–3.60 (m, 2H), 1.76–0.89 (m, 21H). ¹³C NMR (101 MHz, DMSO-d6) δ 159.94, 48.69, 45.21, 33.74, 32.46, 30.14, 28.37, 25.96.

Synthesis of PUU elastomers

PUU-HI was synthesized as follows. PTMEG (20.00 g, 10 mmol) was added into a 250-mL flask and stirred under vacuum at 120 °C for 2 h to remove residual moisture and then cooled to 80 °C. An N, N-dimethylformamide (DMF) solution containing IPDI (4.44 g, 20.0 mmol) and DBTDL (70 µL) was added. The reaction was conducted for 3 h to obtain the prepolymer. After reducing the temperature to 40 °C, the chain extenders of IPDAA (1.43 g, 5 mmol) an in 5 mL DMF were added into the prepolymer solution and stirred for 3 h. Subsequently, the chain extenders of HMDAA (1.63 g, 5 mmol) an in 15 mL DMF were added into the prepolymer solution and stirred for 15 h. A viscous and transparent solution of PUU-HI was finally acquired. The whole reaction was conducted under N₂ atmosphere. The solution was placed in a glass dish and under vacuum environment at room temperature to remove bubbles. Finally, the product was fully reacted in a blast oven at 80 °C for 24 h to obtain a film of PUU-HI.

Characterization

The ¹H-NMR spectra were measured using a Bruker AVANCE III 400 MHz. The solution was deuterated DMSO or CHCl₃, and the internal reference was tetramethylsilane (TMS). The molecular weight of the elastomers was determined by GPC (Hitachi L-2490Hitachi) with the mobile phase DMF at a flow rate of 1 mL min⁻¹. The sample concentration was controlled at 5 mg mL⁻¹. UV–vis transmittance spectra were recorded on a UV-2550 Spectrophotometer (Shimadzu). ATR-FTIR spectra were taken on a Bruker Tensor II spectrophotometer with a scanning range of 4000–500 cm⁻¹ (32 scans and a spectral resolution of 4 cm⁻¹). XRD were conducted on an XRD-D8 FOCUS (Bruker, Germany) with Cu Kα radiation, with data ranging from 10° to 70° and a scanning speed of 10° min⁻¹. DMA was carried out on the DMA Q800 (TA Instruments, USA) under the tension mode (temperature range of −100 to 100 °C and a heating rate of 10 °C min⁻¹). Rectangular elastomer samples (length, width, thickness: 20.0 × 4.0 × 0.4 mm) were tested at a strain of 0.1% and a frequency of 1 Hz. TG was performed to realize the thermal stability of the polyurethane elastomers on a TA Q500 instrument. The temperature was heated from room temperature to 800 °C at a heating rate of 10 °C min⁻¹ under a N₂ atmosphere. Rheology experiments were performed on a Micro Infrared combined rheometer (MAS60) using the disk-type specimen with a size of about 20 ± 1 mm in diameter and 1 ± 0.05 mm in thickness. The temperature sweep experiments were conducted with a temperature range of 30–150 °C with a strain of 1%, a normal force of 1 N, and a frequency of 1 Hz. The frequency sweep experiments were performed over a frequency range of 0.1 to 100 rad s⁻¹ at temperatures from 30 to 150 °C. AFM was performed on SPM9600 (Shimadzu) to obtain phase and height images of elastomers.

Solid-state NMR (SSNMR) measurements

High-field solid-state NMR (SSNMR) experiments were performed on a Bruker AVANCE NEO wide-bore (89 mm) NMR spectrometer at room temperature (25 °C) and operating at frequencies of 400.18 and 100.6 MHz for ¹H and ¹³C, respectively. Conventional 4 mm double-resonance CP/MAS (cross-polarization/magic-angle spinning) probehead was used, and samples with a volume of 52 μl were placed in a zirconia rotor. The MAS was automatically controlled with a speed controller for all one and two-dimensional (1D and 2D) experiments; TOSS (total suppression of spinning sidebands)⁶³ was adopted during ¹³C signal acquisition. The ¹H and ¹³C chemical shifts were referenced to external TMS (0 ppm) and hexamethylbenzene (HMB, 17.3 ppm for CH₃).

Small-angle X-ray scattering (SAXS) measurements and wide-angle X-ray diffraction (WAXD) measurements

SAXS measurements were conducted on a Xeuss system with a multichamber-focused Cu Kα X-ray source (NanoSTAR, Bruker-AXS, Germany, λ = 0.154 nm). The gap from the elastomer to the detector was 1035 nm. 1D SAXS image was acquired over an exposure duration of 10 min using a Vantec 2K detector with 2048 ×  2048 pixels (resolution: 68 × 68 μm²). The 1D SAXS curves were integrated by averaging azimuthally 360° of 2D SAXS patterns via the Fit2D software. Bragg’s law calculated the periodicity (L):

$$L=\frac{2\pi }{q_{\max }}$$ 1

where q_max refers to the peak location of the 1D SAXS curve.

The 2D SAXS patterns were reduced to 1D azimuthal profiles $\mathrm{I}\left(\mathrm{\Phi }\right)$ using the XSACT software. A Q crown was defined with a radial range of $Q_{\min }=0.02{\ddot{\text{A}}}^{-1}$ to ${\mathrm{Q}}_{\max }=0.15{\ddot{\text{A}}}^{-1}$, which covers the principal scattering peak of all samples. An azimuthal sector from ${40}^{\circ }$ to ${50}^{\circ }$ was applied to improve the signal-to-noise ratio, with inversion symmetry enabled to average symmetric angles.

Thus, the azimuthal intensity was obtained by integrating over the finite Q-range:

$$I\left(\mathrm{\Phi }\right)={\int }_{Q_{\min }}^{Q_{\max }}I\left(Q,\mathrm{\Phi }\right)dQ$$ 2

WAXD measurements were also carried out on the NanoSTAR after changing the accessory equipment. WAXD patterns acquired in the center of the sample were collected within 600 s at a distance of 64.0 mm from the elastomer to the detector. All the WAXD patterns were corrected for background using the standard procedure. The 1D WAXD curves were integrated within −90° of 2D WAXD patterns along the vertical direction via the Fit2D software.

Tensile tests

The tensile tests were conducted using a universal testing machine (UTM6503, China Suns) at 25 °C. The size of the dumbbell-shaped elastomer sample is 10.0 × 4.0 × 0.3–0.5 mm. Each kind of sample was tested with at least five individual tensile tests at diversified stretching speeds. The toughness (τ) of an elastic specimen is determined by the area integral below the engineering stress–strain curve.

True stress (σ_t) and true strain (ε_t) are calculated from the engineering stress–strain curves measured at the stretching speed of 50 mm min⁻¹ using the following equation:

$${\sigma }_t=\sigma *\frac{L}{L_0}=\sigma *\left(\epsilon +1\right)$$ 3

$${\epsilon }_t={\int }_{L_0}^L\frac{dL}{L}=\ln \frac{L}{L_0}=\ln \left(\epsilon +1\right)$$ 4

where σ is the engineering stress, L is the instantaneous length of the sample during deformation, L₀ is the length of the sample before testing, ε is the engineering strain.

The toughness (τ) of the elastomer samples is defined by integrating the area under engineering stress–strain curves (σ–ε), measured at the stretching speed of 50 mm min⁻¹, according to the following equation:

$$\tau ={\int }_{\epsilon =0}^{ϵ={\epsilon }_{\max }}\sigma d\epsilon$$ 5

where σ and ε are the engineering stress and engineering strain respectively, and ε_max is the elongation at the break of the elastomer samples.

For the cyclic tensile test, both loading and unloading processes were performed at a strain rate of 50 mm min⁻¹ at room temperature. For the cyclic tensile tests with increasing strain amplitudes, a single specimen was subjected to consecutive loading-unloading cycles at progressively increasing strain levels (e.g., 100%, 200%, and 300%, etc.). Only one cycle was performed at each strain amplitude, and the specimen was continuously tested without any waiting or recovery period between consecutive amplitudes. The reported curves represent the first cycle at each strain level, i.e., non-steady-state cycles, to evaluate the progressive energy dissipation behavior of the material under sequential deformation. In each cycle, the sample was stretched to a strain of 100% and then relaxed at room temperature for a certain waiting time (5 min, 30 min, 2 h, 4 h) prior to the subsequent loading process.

Measurement of the fracture energy

Fracture energy was measured in terms of a tensile test using the single-edge notched sample (length of the crack = 1 mm). The notched and unnotched specimens (gauge length = 10 mm, width = 5 mm, and thickness = 0.4–0.6 mm) were both tested at a crosshead speed of 3 mm min⁻¹, and the fracture energy was calculated from:

$$Fractureenergy=\frac{6wc}{\sqrt{{\lambda }_c}}$$ 6

where c represents the crack length (i.e., 1 mm), λ_c elongation-at-break of the notched sample, w the strain energy estimated by integration of the tensile stress–strain curve of the unnotched specimen until λ_c.

Trouser tear tests

Trouser tear tests were performed on the same universal tester at room temperature under the crosshead speed of 50 mm min⁻¹, using the rectangular sample (40 mm × 20 mm × 0.4–0.6 mm) with an initial crack (crack length = 20 mm). The two legs of the sample were tightly clamped to prevent the sample from slipping during the test. The tear energy was given by:

$$Tearenergy=\frac{{\int }_0^{L}Fdl}{d\cdot L_{bulk}}$$ 7

where ${\int }_0^{L}Fdl$ represents the integrated area under the curve of load (F) versus displacement (L), d sample thickness and $L_{bulk}$ (i.e., 20 mm) the tear length.

To evaluate the recovery capacity of elastomers, different methods were used. One method was to retest the fractured sample after 10 h of contact at 50 °C, and the other was to dissolve the debris in a solvent and cure it back into a film. The recycled sample was further measured by the tensile test. The recycled efficiency or healing efficiency can be calculated by the ratio of toughness before and after recovery, as the following equation:

$$\eta =\frac{{\tau }_0}{{\tau }_n}\times 100\%$$ 8

where η is the recycled efficiency or healing efficiency, τ₀ refers the original toughness of the elastomer, and τ_n is the toughness at the nth recycling time.

Dynamic mechanical analyses (DMA)

DMA were performed on the dynamic mechanical analyzer TA-DMA850 using tension mode under 1 Hz at a heating rate of 5 °C min⁻¹ under a nitrogen atmosphere. The temperature was ramped from −90 to 150 °C.

TENG performance tests

Performance tests of the TENG were measured by the electrometer (Keithley 6514). TENG was driven by a linear motor (QB-300-NM).

Gel permeation chromatography (GPC)

The GPC measurements was conducted on a Hitachi GPC system with Shodex columns (5000−5 K, 400−0.5 K, and 5−0.15 K), a Hitachi L-2130 HPLC pump, a Hitachi L-2350 column oven, and a Hitachi L-2490 refractive index detector. DMF as the elution solvent, and Poly (methyl methacrylate) as the standard.

Author contributions

R. Li prepared the samples, performed characterization experiments and analysis, visualized the data, and prepared the figures. S. Pan performed part of the characterization and analysis. Z. Liu conducted the SAXS experiments and analysis. S. Chen assisted with the characterization experiments and analysis, performed the DMA experiments, visualized the data, and wrote the initial draft of the manuscript. P. Sun supported the experiments, assisted with data analysis, and provided funding. F. Wang performed the molecular dynamics simulations and solid-state NMR experiments and analysis. F. Wang supervised the study, conceived the concept, designed the experiments and analyses, and revised the manuscript.

Peer review

Peer review information

Nature Communications thanks Alexey V. Lyulin and the other anonymous reviewers for their contribution to the peer review of this work. [A peer review file is available].

Funding

P. Sun discloses support for the research of this work from the National Natural Science Foundation of China (grant number 52573082).

Data availability

The Source data generated in this study, including spectroscopic, computational, and mechanical data, have been deposited in the Figshare database under accession code (https://doi.org/10.6084/m9.figshare.31843183). All other data supporting the findings of this study are available in the article and its Supplementary files. All data are available from the corresponding author upon request.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41467-026-77193-w.