White structure flaking failure in bearings under rolling contact fatigue

Abstract

White structure flaking (WSF) as a premature wear failure mode in steel rolling element bearings is caused by white etching cracks (WECs) and perhaps butterflies formed in the ~1 mm zone beneath the contact surface under rolling contact fatigue (RCF). WECs are branching crack systems typically several millimetres in length that have a microstructural change called ‘white etching area’ (WEA) associated with the crack. Butterflies are smaller cracks initiating at material defects and impurities that form WEA wings that revolve around their initiators. Hydrogen diffusion into the bearing steel during service and transient operating conditions have been suggested as drivers of white etching features (butterflies, WEA and WECs). However the initiation and propagation mechanisms as well as the thresholds for WEC formation are not well understood. This is due to the difficulties of creating WECs repeatedly under laboratory conditions and the lack of a method established for mapping WECs in detail or 3 dimensions as typically only limited metallographic analyses are conducted over several cross-sections. A series of RCF tests have been conducted in this study to investigate the formation drivers and formation mechanisms of WECs using a two-roller RCF machine. WECs were successfully created in hydrogen charged 100Cr6 martensitic steel rollers under low-moderate concentrations of diffusible hydrogen (~1 ppm) and service realistic loading conditions (Pmax 1.5 – 2 GPa). However, only butterflies were formed under transient conditions with non-hydrogen charged rollers. One such butterfly was analysed in detail to further understanding of crack formation mechanisms and carbide dissolution as part of the WEA microstructural change. Based on the evidence obtained from the SEM, FIB tomography and STEM/TEM analysis, a void/cavity coalescence theory for initial butterfly crack formation and iron chromium carbide dissolution as part of the WEA formation mechanism is proposed. Metallography was extensively used in this project to view cross-sections of the wear zones subject to RCF. A metallographic serial sectioning technique was established to quantitatively map wear zones for the first time. Mapping WECs in their entirety and 3D modelling revealed the 3-dimensional morphology and orientation of WECs and maximised detection of possible WEC initiators. This study has for the first time quantitatively investigated the influence of diffusible hydrogen, load and rolling cycles on white etching feature formation and the thresholds of formation. The hydrogen charged tests showed that the formation of butterflies was independent of the concentration of diffusible hydrogen with the test parameters used, but dependent on contact pressure and number of rolling cycles up to a threshold. WEC formation thresholds were found at certain values of the concentration of diffusible hydrogen, contact pressure and number of rolling cycles. Extensive serial sectioning and 3D modelling of WECs also demonstrated that the orientation of WECs differed depending on the sectioning direction. It was found that the vast majority of WECs were contained in the subsurface wear zone and did not make any connection with the surface, thus dismissing surface initiation. The WECs often interacted with inclusions that were judged to be crack initiators and evidence was found that butterfly cracks could propagate to form WECs. The white etching features initiated predominately at short sulfide type inclusions, small globular manganese sulfide oxide inclusions and small globular oxide inclusions. Therefore strong evidence was observed for a subsurface initiation mechanism of WECs from non-metallic inclusions. A comparison of the WEC formations in the hydrogen charged two-roller tests was made with serial sectioning investigations of WEC formation in wind turbine gearbox bearings obtained from the field and those tested on a large-scale transient test rig (non-hydrogen charged). This was performed to understand if a difference in the WEC initiation and propagation mechanism occurs under the differing conditions. The comparison showed correlation between the WEC formation mechanisms as a high number of inclusions interacted with the WECs that were judged to be crack initiators and small/short sized inclusions predominated as the crack initiators. Therefore based on the serial sectioning analysis across various test specimens and bearings it is proposed that one mechanism of WEC formation is due to multiple linking of extended butterflies or small WECs in the subsurface to form larger WEC networks that eventually propagate to the surface resulting in WSF. The data also suggests that steel cleanliness standards analysing inclusion density (as opposed to maximum inclusion lengths) are more relevant in understanding butterfly/WEC initiation. However steel cleanliness standards used should record inclusions that are only a couple of micrometer’s in length/diameter.

More information

Identifiers

ORCID for L. Wang: orcid.org/0000-0002-2894-6784

Catalogue record

Export record