A B S T R A C T When the fatigue life N f of a specimen of 10 mm in thickness is longer than 10 8 cycles, the average fatigue crack growth rate is much less than the lattice spacing (~0.1 Å or 0.01 nm) that is 10 −11 to 10 −12 m/cycle. In the early stage of the fatigue process, the crack growth rate should be much less than the average growth rate, and accordingly we cannot assume that crack growth occurs cycle by cycle.In this paper, possible mechanisms for extremely high cycle fatigue are discussed. Of some possible mechanisms, a special focus was put on a newly found particular fatigue fracture morphology in the vicinity of the fracture origin (non-metallic inclusions) of a heat-treated alloy steel, SCM435, which was tested to NÁ10 8 . The particular morphology observed by SEM and AFM was presumed to be influenced by the hydrogen around inclusions. The predictions of the fatigue limit by the ǰarea parameter model are~10% unconservative for a fatigue life of N f =~10 8 , though it successfully predicts the conventional fatigue limit defined for N=10 7 . Thus, the fatigue failure for NÁ10 8 is presumed to be caused by a mechanism which induces breaking or releasing of the fatigue crack closure phenomenon in small cracks.In the vicinity of a non-metallic inclusion at the fracture origin, a dark area was always observed inside the fish-eye mark for those specimens with a long fatigue life. Specimens with a short fatigue life of N f =~10 5 do not have such a dark area in the fish-eye mark. SEM and AFM observations revealed that the dark area has a rough surface quite different from the usual fatigue fracture surface in a martensite lath structure.Considering the high sensitivity of high-strength steels to a hydrogen environment and the high hydrogen content around inclusions, it may be concluded that the extremely high cycle fatigue failure of high-strength steels from non-metallic inclusions is caused by environmental effects, e.g. hydrogen embrittlement coupled with fatigue.

The fracture surfaces of specimens of a heat‐treated hard steel, namely Cr–Mo steel SCM435, which failed in the regime of N = 105 to 5 × 108 cycles, were investigated by optical microscopy and scanning electron microscopy (SEM). Specimens having a longer fatigue life had a particular morphology beside the inclusion at the fracture origin. The particular morphology looked optically dark when observed by an optical microscope and it was named the optically dark area (ODA). The ODA looks a rough area when observed by SEM and atomic force microscope (AFM). The relative size of the ODA to the size of the inclusion at the fracture origin increases with increase in fatigue life. Thus, the ODA is considered to have a crucial role in the mechanism of superlong fatigue failure. It has been assumed that the ODA is made by the cyclic fatigue stress and the synergetic effect of the hydrogen which is trapped by the inclusion at the fracture origin. To verify this hypothesis, in addition to conventionally heat‐treated specimens (specimen QT, i.e. quenched and tempered), specimens annealed at 300 °C in a vacuum (specimen VA) and the specimens quenched in a vacuum (specimen VQ) were prepared to remove the hydrogen trapped by inclusions. The specimens VA and VQ, had a much smaller ODA than the specimen QT. Some other evidence of the influence of hydrogen on superlong fatigue failure are also presented. Thus, it is concluded that the hydrogen trapped by inclusions is a crucial factor which causes the superlong fatigue failure of high strength steels.

High cycle fatigue fracture surfaces of specimens in which failure was initiated at a subsurface inclusion were investigated by atomic force microscopy and by scanning electron microscopy. The surface roughness Ra increased with radial distance from the fracture origin (inclusion) under constant amplitude tension–compression fatigue, and the approximate relationship: Ra ≅ CΔK 2I holds. At the border of a fish‐eye there is a stretched zone. Dimple patterns and intergranular fracture morphologies are present outside the border of the fish‐eye. The height of the stretch zone is approximately a constant value around the periphery of the fish‐eye. If we assume that a fatigue crack grows cycle‐by‐cycle from the edge of the optically dark area (ODA) outside the inclusion at the fracture origin to the border of the fish‐eye, we can correlate the crack growth rate da/dN, stress intensity factor range ΔKI and Ra for SCM435 steel by the equation and by da/dN proportional to the parameter Ra . Integrating the crack growth rate equation, the crack propagation period Np2 consumed from the edge of the ODA to the border of the fish‐eye can be estimated for the specimens which failed at Nf > 107. Values of Np2 were estimated to be ∼1.0 × 106 for the specimens which failed at Nf ≅ 5 × 108. It follows that the fatigue life in the regime of Nf >107 is mostly spent in crack initiation and discrete crack growth inside the ODA.

No abstract

No abstract