Mechano-responsive hydrogen-bonding array of thermoplastic polyurethane elastomer captures both strength and self-healing

Abstract

Self-repairable materials strive to emulate curable and resilient biological tissue; however, their performance is currently insufficient for commercialization purposes because mending and toughening are mutually exclusive. Herein, we report a carbonate-type thermoplastic polyurethane elastomer that self-heals at 35 °C and exhibits a tensile strength of 43 MPa; this elastomer is as strong as the soles used in footwear. Distinctively, it has abundant carbonyl groups in soft-segments and is fully amorphous with negligible phase separation due to poor hard-segment stacking. It operates in dual mechano-responsive mode through a reversible disorder-to-order transition of its hydrogen-bonding array; it heals when static and toughens when dynamic. In static mode, non-crystalline hard segments promote the dynamic exchange of disordered carbonyl hydrogen-bonds for self-healing. The amorphous phase forms stiff crystals when stretched through a transition that orders inter-chain hydrogen bonding. The phase and strain fully return to the pre-stressed state after release to repeat the healing process.

Fig. 1. Chemical structures and mechanical properties of self-healable TPUs.

a Chemical structures and b tensile stress–strain curves for the three TPUs: C-IP-SS, E-IP-SS, and commercial Es-MD. Tensile curves of cut and rejoined C-IP-SS with different healing times at 35 °C are included. c Comparing mechanical properties, such as elongation at break, ultimate tensile strength, and toughness, for virgin and healed TPUs. The value for the non-self-healable Es-MD is given as a point in each panel. d Photographic images of cut and healed C-IP-SS films before (left) and after stretching up to 400% (right) (scale bar: 1 cm). e C-IP-SS film (thickness: 0.3 cm) cut in half, rejoined, and healed for 48 h at 35 °C, followed by a 10-kg weight-lifting demonstration (scale bar: 5 cm). f Photographic images of a twisted C-IP-SS film (thickness: 0.1 cm) after healing at 25 °C for 1 min (scale bar: 1 cm).

Fig. 2. Ashby plot of ultimate tensile strength versus self-healing temperature of C-IP-SS and other elastomers reported in literature.

The red star indicates the properties of C-IP-SS. The blue circles indicate the properties of the reported autonomously self-healing elastomers without external stimulus. The yellow triangles indicate the properties of the reported elastomers that need heat for healing.

Fig. 3. Mechano-responsive dual-mode of C-IP-SS.

a Variations in tensile modulus versus strain obtained from the first derivatives of the tensile stress–strain curves of the TPUs (Fig. 1b). b Macroscopic structural changes undergone by the TPUs upon uniaxial stretching to 400% (scale bar: 6 mm). c Optical microscopy images of unstretched and 400%-stretched C-IP-SS specimens. d Schematic illustration of mechano-responsive dual-mode operation through the reversible disorder-to-order transition of the H-bond array: self-healing in the static state and toughening in the dynamic state.

Fig. 4. FT-IR analysis of C-IP-SS in the static and dynamic states.

a The carbonyl region of the FT-IR spectrum of C-IP-SS with peak deconvolution. b 2D gradient FT-IR map with respect to stretching percentage. The red contour lines represent positive values of dA/dE (the first derivative of the absorbance (A) as a function of the extension % (E)); i.e., an increasing trend in absorbance as a function of the degree of extension, and vice versa, for the gray contour lines. c Schematic of the mechano-responsive changes in chain alignment and the associated H-bond arrays for C-IP-SS.

Fig. 5. Rheological evidence for the self-healing mechanism.

a Rheological parameters for the three TPUs in static mode: complex viscosities (η*_0.05) were obtained from viscosity curves at 0.05 rad s⁻¹ (Supplementary Fig. 13a). Yield stresses were calculated from Casson plots (Supplementary Fig. S13b) and Cole–Cole plot slopes were obtained at 25 °C (Supplementary Fig. S13c). b Variations in loss tangent (tan δ) over the 25–95 °C temperature range at 1 rad s⁻¹. The dashed line indicates tan δ = 1. c Master curves of storage (G′) and loss (G″) moduli for C-IP-SS and E-IP-SS. The reference temperature is 25 °C. d Illustration depicting the representative G′ and G″ master curves. e Relaxation times for chain flow transition (τ_f) and segmental motion (τ_s) for C-IP-SS and E-IP-SS at 25 °C. The τ_f values were determined from the reciprocal values of the crossover frequencies of the G′ and G″ master curves (Fig. 5c), and the τ_s values were obtained from relaxation curves at 0.05 rad s⁻¹ that were calculated using Eq. (3) (Supplementary Fig. 14).

Fig. 6. X-ray and thermodynamic analyses of strain-induced phase transition.

a, 2D WAXS images of three TPUs at stretching ratios of 0%, 200%, and 400%. The WAXS results of 400%-stretched samples and that were held for 72 h are also included to trace further structural evolution. b, 1D WAXS profiles of C-IP-SS at various degrees of stretching. c Stress–strain curves of C-IP-SS at various temperatures in the −30 to 40 °C range. d Force versus temperature plots at various strains based on the stress–strain curves.

Fig. 7. Biocompatibility test of C-IP-SS.

a Experimental procedure for in vivo histocompatibility testing: rat subcutaneous connective tissues with negative control (NC, HDPE), positive control (PC, DMSO-containing polyurethane), and the C-IP-SS film. b Representative histopathologic tissue images after 12 weeks of healing. Arrows indicate inflammatory cells (red: polymorphonuclear cells; black: lymphocytes).