Efeito da reversão da martensita induzida por deformação sobre o comportamento e eletroquímico do aço inoxidável lean duplex 2304

Abstract

Lean duplex stainless steels (AILD) are cost-effective alternatives to traditional duplex grades such as 2205 and to austenitic steels like 304L and 316L, comprising a biphasic ferrite–austenite microstructure with relatively low nickel and molybdenum contents. Under severe plastic deformation, metastable austenite can partially be transformed into strain-induced martensite (α′), which can be reversed by thermal annealing. This work investigated the effects of the formation and partial reversion of martensite on the microstructural, mechanical, fractographic, and electrochemical behavior of AILD 2304, cold rolled to 74% thickness reduction and subsequently subjected to isochronal annealing between 400 °C and 1100 °C for 30 minutes. Microstructural analysis was conducted via XRD, SEM, TEM, and EBSD. Mechanical properties were evaluated by tensile and microhardness testing, while electrochemical behavior was studied by potentiodynamic polarization, DL-EPR using a syringe cell, and statistical modeling of breakdown potentials. Pit nucleation sites were investigated by potentiostatic pulse experiments. Results showed that partial martensite reversion begins between 500 °C and 550 °C, accompanied by the formation of a nanostructure rich in stacking faults, twins, and dislocations. Aging at 400 °C and 500 °C led to increased hardness and strength but also to loss of ductility and the occurrence of brittle fracture. Cold rolling significantly enhanced pitting corrosion resistance, whereas annealing between 400 °C and 600 °C lowered the passive film breakdown potential and promoted pit nucleation at α′ martensite/austenite interfaces, with increased pit density. Sensitization was detected from 500 °C onward, even in the absence of visible carbide precipitation, associated with local elemental redistribution. A bimodal behavior involving both pitting and transpassivity was observed in as-received and 400 °C-annealed specimens. Increasing the exposed area raised the probability of pit initiation. Statistical modeling using a mixed distribution (Gumbel + normal) enabled prediction of Eb variation with surface area, emphasizing the importance of probabilistic approaches and sufficient repeatability in pitting corrosion studies, as well as enabling extrapolation of Eb to larger areas.