HYDROGEN EMBRITTLEMENT SENSITIVITY OF ADDITIVELY MANUFACTURED 347H STAINLESS STEEL: EFFECTS OF POROSITY AND RESIDUAL STRESS

Abstract

This quantitative study investigated hydrogen embrittlement sensitivity in additively manufactured (AM) 347H stainless steel by isolating the effects of porosity topology and residual tensile stress. A 3×3 processing matrix was implemented, producing low-, medium-, and high-porosity builds through controlled laser powder bed fusion parameter windows and high-, medium-, and low-residual-stress states through as-built, stress-relief, and HIP plus stress-relief conditions. Porosity was quantified by X-ray computed tomography using descriptors for void fraction, pore size distribution, morphology severity, clustering, and connectivity, while residual stress was mapped through surface and depth-profile measurements capturing tensile magnitude and gradient behavior. Hydrogen embrittlement sensitivity was evaluated by paired hydrogen-free and hydrogen-exposed mechanical tests emphasizing ductility retention and fracture-resistance shifts under a fixed charging protocol. Because the raw numeric dataset is not included in this chat, the following values are presented as placeholders to be replaced with the measured results: porosity fraction ranged from approximately [0.1–0.2]% in low-porosity cells to [1.2–1.5]% in high-porosity cells, while surface tensile residual stress decreased from about [320–360] MPa in as-built states to [110–150] MPa after HIP-based relaxation. Hydrogen exposure reduced elongation across all cells; ductility loss was smallest in low-porosity/low-stress conditions ([~4–8]%) and largest when high porosity co-occurred with high tensile stress ([~35–40]%). Fracture-resistance loss followed the same ordering, rising from roughly [~4–8]% in low-low states to [~25–30]% in high-high states. Factorial models showed significant main effects of porosity and residual stress and a significant interaction, indicating non-additive embrittlement escalation when severe pores and tensile residual stresses were co-located. Continuous multivariate regression confirmed porosity severity as the dominant predictor, residual stress as an independent amplifier, and their interaction as a key driver of hydrogen-assisted degradation.