3.2. Fracture surface
    The SCC was broken open and the fracture surface was examined. There were some crack branches close to the pipe outer surface because there were some small facets at the fracture surface. Most of the SCC surface was covered with a visible back film. Parts of the fracture surface were brown in colour, these were closer to the original pipe surface. This is interpreted as being due to easier access of the external environment along the more open crack before the position F in Fig. 3.
   There were several crack initiation sites at this fracture surface. The small facets close to the pipe surface showed that the cracks initiated and propagated individually first. The cracks coalesced when they grew deeper. The structure of the oxide film at the different positions on the fracture surface was characterised and a typical topography is shown in Fig. 4. The oxide film at the edge formed for a longer time than that close to the crack tip. Most of the crack surface was covered by oxide film. Fig. 5 showed the structure around the crack tip. There was a narrow region of the crack tip that was not covered by an obvious oxide film. It showed intergranular stress corrosion cracking. The crack tip front was next to a tearing surface produced by the ductile overload as the sample was broken open to reveal the SCC surface for observation.
3.3. Depth profiles
    Depth profiles of the surface film on the SCC fracture surface A1 were analysed by Auger electron spectroscopy (AES), at the edge, middleand tip positions as shown in Fig. 6. The relationship between the thickness of the oxide film and the crack length is shown in Fig. 6(d). The edge had the thickest film because of the longest time for corrosion, and the tip had the thinnest film because of the shortest time exposure to the environment. Due to the resolution of the AES, it is impossible to analyse the thickness of the oxide film right at the crack tip.