Measurement and modeling of hydrogen environment assisted cracking of ultra-high strength steel

Traditional martensitic ultra-high strength steels such as AISI 4340 achieve high yield strength (σ_YS > 1400 MPa) and plane strain fracture toughness (K_IC ﹥MPa-√m) in benign environments, but are highly susceptible to hydrogen assisted cracking (HAC) when exposed to H through processing or service. Numerous studies over several decades have evidenced that hydrogen interacts with segregated impurity atoms along grain boundaries to cause intergranular cracking on these older steels, and the current generation of ultra-high strength steels attempt to address this problem through improved purity. Modem precipitation hardened ultra-high strength AerMet™100 steel (Fe-Co-Ni-Cr-Mo-C) is vacuum forged to limit S and P impurity concentrations well below 0.005 weight %. With optimal heat treatment, AerMet™ 100 has a very high tensile strength (σ_YS > 1700 MPa) due to a uniform distribution of nanoscale carbides in an unrecrystallized martensite matrix, while maintaining high fracture toughness (K_IC > 100 MPa√m) thanks to the absence of cementite and formation of 3 nm thick reverted austenite between martensite laths.

Despite the improvements in strength, toughness and purity, slow rising-extension tests on fully immersed AerMet™ 100 specimens showed susceptibility to severe transgranular hydrogen environment assisted cracking (HEAC) in neutral 3.5% NaCl solution. The threshold stress-intensity for HEAC, K_TH, is reduced to as low as 10% of K_IC and Stage II subcritical crack growth rate, da/dt_II, is up to 0.5 μm/s. Low K_TH and high da/dt_II are produced at potentials substantially cathodic, as well as mildly anodic, to free corrosion. However, a range exists at slightly cathodic potentials (-0.625 to -0.700 V_SCE) where crack growth rate is greatly reduced, consistent with reduced crack tip acidification and low cathodic overpotential for limited H uptake.

Valid HEAC measurements of crack growth rate and threshold are obtained from relatively short tests of 1 to 3 days using the combination of high resolution in-situ crack monitoring and slow. extension rate testing. Considering the specimen design, short crack size (250 to 1000 μm) does not promote unexpectedly severe HEAC. High purity AerMet™ 100 is susceptible to HEAC because martensite boundary trapping and high crack tip stresses strongly enhance H segregation to sites that form a transgranular crack path. Therefore, high purity steels are not necessarily immune from HEAC, especially when H trapping is significant. However, the exact reason for a preferred TG cracking over IG cracking remains unknown. The uniformly dispersed nano-scale M₂C carbides do not enhance the HEAC threshold as previously suggested, and their role in crack growth rate is speculative.

A semi quantitative crack growth rate model was formulated using a simplified one dimensional analysis for stress driven H diffusion to the crack initiation site from the crack tip surface. This model predicts the applied potential dependence of da/dt_II using reasonable input parameters, particularly crack tip H uptake reverse calculated from measured K and a realistic critical distance. The correlation suggests that Stage II da/dt is H diffusion rate limited for all potentials examined, but significant modeling challenges remain.