EBSD analyses on unfailed notched samples

In each of the notched bar samples, two geometrically symmetric notches were made in the gauge length. The creep failure only occurs at one of the notches and leaves the other notch macroscopically intact. This provides an opportunity to study the unfailed notch region to understand the actual failure mode which caused failure on the second notch. The samples were therefore sectioned along the gauge length. In the notched area, large amounts of cavities, cavity linkages and cracks can be observed in SEM examination (see Figure 6.37).

The presence of the notches leads to a stress concentration and accelerates the creep testing process. When one of the notches fails in creep testing, the unfailed one is actually at a near failure state. It has been discussed in Chapter 4 that the locations of creep cavities do not sit on any specific type of grain boundaries (e.g. PAGBs, martensitic lath boundaries, etc.). In some cases, the average size of cavities is even bigger than the width of martensitic lath, and therefore not possible to confirm the relationship between cavity nucleation sites and lath boundaries. EBSD large area mapping on the cracks and cavity linkages at the near failure state can be useful to investigate the crack propagation path selection to see if the cracks follow any specific grain boundaries.

The areas with obvious cracks and cavity linkages in the unfailed notch regions of the three samples were mapped by EBSD. Figure 6.38 and 6.39 show the microcracks and cavity linkages in BM A. The EBSD maps of the cracks developed in the notch region of BM B are shown in Figure 6.40 – 6.44. The cracks in Figure 6.45 – 6.47 are from the notch region of BM C. Ten areas in total were EBSD-mapped in the three samples. It can be summarised from the maps that most cracks propagate sometimes along the PAGBs and sometimes into the PAGs. Since the crack propagation follows a mixed path, inter- and trans-granular, an important conclusion can be drawn that the low ductility failure of G92 steels is not originated from the grain boundary embrittlement or related to theories concerning the failure of grain boundaries.

Figure 6.37 Low magnification SEM images showing the distribution of microcracks and cavities in the unfailed notched regions of (a) BM A, (b) BM B and (c) BM C.

Figure 6.38 Coarse EBSD maps showing the microcrack propagation path in the unfailed notch region of BM A (area 1): (a) IQ (image quality) map, (b) IPF (inversed pole figure) + PAGBs (prior austenite grain boundaries), and (c) non-indexed points + PAGBs.

(a) (b) (c)

Figure 6.39 Coarse EBSD maps showing the cavity linkages in the unfailed notch region of BM A (area 2): (a) IQ map, (b) IPF + PAGBs, and (c) non-indexed points + PAGBs.

(a) (b) (c)

Figure 6.40 Coarse EBSD maps showing the crack propagation path in the unfailed notch region of BM B (area 1): (a) IQ map, (b) IPF + PAGBs, and (c) non-indexed points + PAGBs.

(a) (b) (c)

Figure 6.41 Coarse EBSD maps showing the crack propagation path in the unfailed notch region of BM B (area 2): (a) IQ map, (b) IPF + PAGBs, and (c) non-indexed points + PAGBs.

(a) (b) (c)

Figure 6.42 Coarse EBSD maps showing the crack propagation path in the unfailed notch region of BM B (area 3): (a) IQ map, (b) IPF + PAGBs, and (c) non-indexed points + PAGBs.

(a) (b) (c)

Figure 6.43 Coarse EBSD maps showing the crack propagation path in the unfailed notch region of BM B (area 4): (a) IQ map, (b) IPF + PAGBs, and (c) non-indexed points + PAGBs.

(a) (b) (c)

Figure 6.44 Coarse EBSD maps showing the crack propagation path in the unfailed notch region of BM B (area 5): (a) IQ map, (b) IPF + PAGBs, and (c) non-indexed points + PAGBs.

(a) (b) (c)

Figure 6.45 Coarse EBSD maps showing the crack propagation path in the unfailed notch region of BM C (area 1): (a) IQ map, (b) IPF + PAGBs, and (c) non-indexed points + PAGBs.

(a) (b) (c)

Figure 6.46 Coarse EBSD maps showing the cavity linkages in the unfailed notch region of BM C (area 2): (a) IQ map, (b) IPF + PAGBs, and (c) non-indexed points + PAGBs.

(a) (b) (c)

Figure 6.47 Coarse EBSD maps showing the crack propagation path in the unfailed notch region of BM C (area 3): (a) IQ map, (b) IPF + PAGBs, and (c) non-indexed points + PAGBs.

6.6 Summary

In this chapter, the microstructures of three non-creep tested base metals A, B and C were characterised in detail. Characterisation was also carried out on plain bar and notched bar creep samples in the same materials. The ‘worst’

steel BM A and the ‘best’ steel BM C were selected for comparison in some cases due to their extreme performance in creep testing.

The key inclusion particles, BN, MnS and Al2O3, characterised in the previous chapters were all found in the three steels with different characteristics. BM A is the only steel whose microstructures are strongly affected by the product orientation. In BM A, large amounts of BN / MnS / Al2O3 / reversible ferrite

(a) (b) (c)

pipe, which is believed to be the result of the vertical extrusion process. The heterogeneous distribution of inclusion particles in BM A led to the heterogeneous distribution of creep cavities compared to BM C.

Nb-rich boride particles were found near the inclusion clusters in BM A. This phase was also observed in BM C with very small particle sizes and low quantities. Needle shaped AlN particles were also observed and identified by EDS in BM B due to the high Al content.

BN was quantified in the head sections of plain bar samples of three steels.

The total area percentages of BN in each sample were similar. Although BM B has the highest number of BN, the quantification result may be affected by the presence of AlN particles which appear dark in BSD-SEM examination. BM C has twice as many BN particles with smaller particle sizes than BM A. Linking to the investigation of cavity / BN association in the low stress regions of the notched bar samples, the BN particles in BM C appear to be much more difficult to initiate cavities than BM A. A possible explanation is that there is a critical size of BN particles such that only the particles with a size above the critical value can trigger the cavity nucleation. The initiation of cavities on those particles smaller than the critical value is thought to be more difficult.

EBSD mapping on the unfailed notch regions were performed to investigate the failure mode. Since both intergranular and transgranular crack propagation were observed, no clear information supports grain boundary embrittlement or grain boundary failure theory in G92 steels.

CHAPTER 7 THE EFFECTS OF BN ON THE PRECIPITATION OF MX AND effect on the grain boundary diffusion to suppress microstructural degradation.

However, boron itself is a strong nitride forming element which consumes soluble nitrogen at high temperatures to form BN particles. The detrimental effect of BN inclusions on creep behaviour has been related to the nucleation of creep cavities, which are thought to be one of the important reasons of the reduction of creep ductility and rupture strength. On the other hand, the consumption of free nitrogen also reduces the initial benefit of adding this element to provide a precipitation hardening effect by combining with certain elements such as vanadium and niobium to form a fine distribution of MX particles.

The precipitation of BN therefore was not originally designed to be the case for G92 steels and the open literature focusing on the BN issues in G92 steels is scarce. BN has been shown to be associated with cavity nucleation sites in the previous chapters. However, the effects of BN formation on the other microstructural features have not been addressed systematically. This chapter has therefore investigated the actual dissolution temperature of BN inclusions using a dilatometer to achieve precise temperature control. In addition, the key second phases, MX and M23C6, which are most likely to be affected by BN, were investigated in parallel. The aim of this chapter is to investigate the BN dissolution behaviour as a function of temperature and to provide a quantified understanding of the effects of BN on the precipitation and evolution of MX and M23C6.