Hydrogen embrittlement in high strength fasteners: Comparison between bainitic and tempered martensitic steels

The scope of this work is to explore innovative microstructure typologies, as well as heat treatments, for the manufacturing of high-strength steel fasteners aimed at obtaining a reduced susceptibility to hydrogen embrittlement. The research evaluates the role of microstructure and the notch tensile strength (RNTS) in relation to the geometry tested on hydrogen embrittlement of steels used in fasteners production. High-strength steel fasteners are increasingly used especially in Automotive industry to mechanically join parts together, increase vehicle performance and promote car weight reduction. High-strength steel fasteners, included in property class 12.9 and higher, are characterized by an ultimate tensile strength above 1200 MPa, despite the remarkable mechanical performance they are historically prone to hydrogen embrittlement phenomena. The phenomenon of hydrogen embrittlement is widely investigated, characterized by its unpredictability and its ability to result in brittle fractures. In this study, slow strain rate tensile tests were conducted to investigate and compare the effect of internal hydrogen concentration on the mechanical properties of fasteners with martensitic microstructure in property classes 10.9and 12.9, as well as innovative high strength fasteners with bainitic microstructure in property class 12.9U. The commonly utilized methodologies for assessing hydrogen embrittlement in fasteners are standardized procedures primarily focused on mechanical tests conducted on finished fasteners. They do not effectively differentiate the influence of surface treatment from material selection and fastener design. For this reason, during the present research, a simpler, faster and very promising methodology has been used. This procedure could be the basis for developing a new Standard guideline for prevention of hydrogen delayed fracture. Additionally, SEM fractography has been interpreted according to the well-known HELP and HEDE models.