Multiscale Structural Evolution Governing Fatigue Resistance in Polyurethane Elastomers

Fatigue failure under long-term cyclic loading remains a central challenge in the design of elastomeric biomaterials that must simultaneously achieve high strength, large deformability, and long-term durability. Here, we report a polyurethane elastomer constructed via a multiscale energy-dissipation strategy that establishes a direct structure-fatigue-property relationship. By integrating molecular-scale chemical cross-linking, microscale physical cross-linking, and dynamic hydrogen-bonding interactions into a synergistic hierarchical network, the elastomer exhibits exceptional tensile strength (68.3 MPa), outstanding toughness (374.09 MJ/m3), and an extraordinary load-bearing capability of 4.27 × 105 times its own weight. Remarkably, the material sustains up to 9.5 × 105 loading cycles at >100% strain and delivers a high fatigue threshold of 793.68 J/m2. Multiscale structural characterizations reveal that fatigue degradation is governed by the progressive evolution and eventual disintegration of the spherulitic crystalline structures. Under cyclic loading, soft-segment crystals are first disrupted, leading to a pronounced reduction in crystallinity, followed by the gradual degradation of hard-segment crystals and fibrillar bundle dissociation. These processes are accompanied by weakened microphase separation, attenuated hydrogen bonding, a reduced long period, and diminished transition layer thickness. Bond-level analysis further identifies C–N and, especially, C–O bonds as the most vulnerable sites under cyclic loading. Notably, the material also exhibits excellent biocompatibility and superior resistance to degradation, positioning it as a promising candidate for artificial ligament applications. This work not only provides a general molecular and structural design paradigm for fatigue-resistant elastomers but also offers new mechanistic insights into the microscopic origins of fatigue durability in biofriendly polyurethane systems, enabling their potential applications in artificial ligaments and flexible sensors.