Effects of cycle frequency, waveform, and electrode potential on corrosion-fatigue crack growth in high-strength tempered martensitic steels

2017-03-02T03:08:01Z (GMT) by Knop, Mark
Corrosion fatigue of steels in sea-water is of particular interest in regard to predicting the lives of submarines and offshore structures. However, the focus of previous research has been mainly directed towards offshore structures, which are subjected to wave loading (sinusoidal waveforms around 0.1 Hz). Submarines, on the other hand, are subjected to a diving cycle with a trapezoidal waveform and slow cycle frequencies (<0.001 Hz). Obtaining corrosion-fatigue crack-growth rates for submarines using in-service cycle frequencies is impracticable in the laboratory as the cycle frequencies are so slow. Consequently, a detailed understanding is required of the variables on corrosion-fatigue crack-growth to enable the extrapolation of data. As such, the main focus of the present study was to obtain a detailed understanding of the effects of cycle frequency and waveform on corrosion-fatigue crack-growth rates in sea-water of high-strength steel and weld metal. To achieve the above aims, corrosion-fatigue crack-growth-rate testing was carried out for a range of cycle frequencies (from 5 Hz to about 0.001 Hz), electrode potentials (open-circuit potential to −1400 mV(SCE)), and waveforms (considering the effect of rise, fall and hold times) in aqueous NaCl on tempered martensitic steel and acicular ferrite weld metal (600–900 MPa), with the Kmax value below the sustained-load stress-corrosion-cracking threshold K value. The corrosion-fatigue crack-growth rates were typically obtained at constant ∆K to isolate the effect of ∆K and allow the variable under consideration (e.g. cycle frequency) to be studied in isolation. The resulting fracture surfaces were examined in detail using optical and scanning electron microscopy. The results showed for both cathodically protected and freely corroding steels that the crack-growth-rate versus cycle-frequency behaviour could be well described using sigmoidal curves at a particular constant ∆K. With increases in constant ∆K, there was a systematic upwards and sidewards shift of these sigmoids, which explains the trends of the crack-growth rate observed on plots of da/dN versus ∆K. In terms of the effect of waveform, the rise time had the predominant influence on the crack-growth rate. The fall time and hold times could also enhance the crack-growth rate but only when rise times were short. The results indicated that the rate-controlling process for corrosion fatigue of steel in aqueous environments is the surface-reaction kinetics (generation and adsorption of hydrogen). Under conditions of cathodic protection, the results supported previous modelling studies that hydrogen generated at the side surfaces of specimens enhances the crack-growth rate. The present study showed that the influence of hydrogen generated at the side surfaces depends on the cycle frequency. Fractography of the steel revealed that hydrogen-assisted crack growth was associated with transgranular cleavage-like fractures and intergranular fractures associated with substantial localised plasticity. When all the results in the present study are considered together with previous findings, the corrosion-fatigue crack-growth behaviour of the steels could be best explained in terms of the adsorption-induced dislocation emission (AIDE)/nanovoid coalescence mechanism for hydrogen-assisted cracking.