Microstructure-environment related fatigue deformation-cracking behavior and data-physics driven crack growth life prediction of perfluorosulfonic acid ionomer

Perfluorosulfonic acid (PFSA) ionomers often suffer fatigue failure under complex environmental conditions, limiting their long-term reliability. However, the effects and underlying mechanisms of environmental factors such as hygrothermal conditions and corrosion on their multi-scale constitutive response and fatigue crack growth (FCG) behavior remain poorly understood. To address this gap, we integrate experimental insights with a physics-informed machine learning (PIML) framework for fatigue life prediction. In-situ experiments and microstructural characterization were performed to quantify the effects of temperature, humidity, and corrosive environments on the viscoelastic response and fatigue crack growth of PFSA ionomer. Results reveal that under ambient conditions, PFSA exhibits typical viscoelastic behavior, explained by a proposed three-level hierarchical structural model. Hygrothermal exposure leads to significant thermal softening and network disruption through hydrogen bonding, and water cluster interactions, whearas chemical corrosion induces side-chain damage and subsequent macroscopic structural destruction. These environmental degradations markedly accelerate FCG rates and alter fracture morphologies. Leveraging these insights, a dual-network Physics-Informed Machine Learning model (PIML-PR) was developed to predict FCG behavior under coupled environmental and mechanical conditions. By fusing physically derived parameters with environmental descriptors, the model achieves high predictive accuracy (≈98 %) in capturing environment-sensitive FCG behavior. This integrated experimental–computational framework provides a pathway for physics-informed lifetime assessment of polymer electrolyte membranes under service-relevant conditions.