Quantitative evaluation of hydrogen segregation at tilt grain boundaries and dislocations in Fe–Cr–Ni alloy

Hydrogen interactions with microstructural defects play a critical role in hydrogen-induced degradation of austenitic stainless steels, yet a quantitative description of hydrogen solubility across defects of different dimensionalities remains limited. This work combines a machine learning interatomic potential with grand canonical Monte Carlo simulations to systematically evaluate hydrogen solubility in the Fe0.7Cr0.19Ni0.11 solid solution, spanning the bulk phase, two-dimensional grain boundaries, and one-dimensional edge dislocations. Hydrogen incorporation is quantified under realistic temperature and pressure conditions by explicitly accounting for the statistical distribution of hydrogen binding energies associated with local atomic environments. This work introduces the concept of excess hydrogen solubility to provide a unified, quantitative measure of hydrogen segregation relative to the bulk phase. These results reveal that twist grain boundaries exhibit negative excess solubility, whereas tilt grain boundaries and edge dislocations act as effective hydrogen traps due to defect-induced structural distortion and strain fields. The observed trends are governed by the thermodynamic balance between hydrogen chemical potential and defect specific binding energy distributions, rather than defect dimensionality alone. This work establishes a transferable thermodynamic framework for predicting hydrogen defect interactions in chemically complex alloys, with direct implications for understanding and mitigating hydrogen-induced degradation.