PSI - Issue 42

Margot Pinson et al. / Procedia Structural Integrity 42 (2022) 471–479 Margot Pinson / Structural Integrity Procedia 00 (2019) 000–000

474

4

Figure 2 shows a SEM image and EBSD maps of the final microstructure of the tempered lightweight bearing steel. The EBSD phase map indicates the formation of retained austenite after quenching. However, XRD measurements only detect about 0.7 vol.% of fcc thus the points indexed as fcc are most likely linked to measurement points with a low confidence index. The PAG size is 24.1 µm and the final bulk hardness is 785 HV, which is very similar to the bulk hardness of the industrial bearing steel. After tempering, no clear kappa carbides are detected with SEM even though a special soft annealing treatment was performed to introduce kappa carbides in the microstructure. However, the TEM analysis in Figure 3 shows small rod-like carbides present in the microstructure after quenching. Another important remark is that a high amount of cracks is present at the prior austenitic grain boundaries (PAGBs) after quenching (cf. Figure 2(A)). These so-called quench cracks arise from the high stresses introduced during the quenching step and they can deteriorate the coherency of the steel matrix (Pinson et al., 2021).

Figure 2: (A) SEM image, (B) phase map and (C) IPF map measured by EBSD of the tempered Fe-8Al-1.1C material.

Figure 3: TEM analysis of the quenched Fe-8Al-1.1C samples, kappa carbides are indicated by white arrows.

3.2 Interaction between H and bearing materials Although both materials have a very similar bulk hardness, their interaction with H is completely different. The melt extraction results in Figure 4 show that the industrial 100Cr6 bearing steel has a much higher H saturation limit. The H diffusion coefficient can also be calculated from these saturation curves, by fitting the experimental data to an analytical solution of the second law of Fick (Claeys et al., 2020; Olden et al., 2008). This results in a H diffusion coefficient of 4.9E-11 m 2 /s and 7.7E-11 m 2 /s for the industrial and the lightweight bearing material, respectively. The differences in H saturation value and H diffusion coefficient can be linked to the differences in H trapping sites between the two materials. A previous study by the same authors, has shown that H in martensitic steels is mainly situated at the dislocations, the high angle grain boundaries (HAGBs) and the carbide interface (Pinson et al., 2022). The reason why the saturation limit of the 100Cr6 sample is much higher than that of Fe-8Al-1.1C can be attributed to the differences in these three microstructural features. Firstly, the 100Cr6 steel is tempered at a lower temperature (160°C) than the Fe-8Al-1.1C steel (250°C). Thus, less dislocations are assumed to be annihilated in the 100Cr6 samples, which can also be deduced from the slightly higher hardness of the 100Cr6 steel compared to the Fe-8Al-