|
The crack
surface is expected to corrode with time, so that the surface layer
thickness increases with time, and the surface layer becomes more obvious.
This phenomenom is routinely used in the examination of fracture surfaces
caused by fatigue. A thick obvious surface layer is to be expected for a
dormant crack that has been dormant for a significant period, and has been
exposed to a corrosive environment during that period of dormancy.
However,a surface layer indistinguishable from that on an active crack
could be expected for a dormant crack that has been dormant for a
significant period, but has not been exposed to a corrosive environment
during that period of dormancy. Conversely, a thick obvious surface layer
is to be expected for an active crack that has been exposed to a corrosive
environment during that period between active cracking and analysis in the
laboratory.
Two pipe segments were received at UQ. These segments,
designated A1 and A2, measured about 0.3 0.3 m. The samples were from the
same excavation, with sample A1 about 10 m upstream. The pipeline had been
in service for a considerable time. The SCC was found during a dig-up and
inspection program and documented in situ. The pipe was coated with
luxepoxy and reburied until the pipe segments were extracted for the
present analysis. After cutting out the segments, the luxepoxy was removed
by grit blasting, and the samples sent to UQ.
At UQ, the surfaces of segments A1 and A2 were examined
visually, magnetic particle inspection was used to detect stress corrosion
cracks in these two pipe segments and the detected cracks were
photographed. The SCC were subjected to examination using: (1) detailed
fractography involving scanning electron microscopy (SEM) and (2) depth
profiling using Auger Electron Spectroscopy (AES).
The examination had the following two aims.
1. From the two coupons cut from the pipeline containing the service SCC,
determine whether the SCC was active or dormant.
2. Evaluate the possibility of taking small cuts from the surface of the
pipeline in-situ that would be suitable for carrying out the above
assessment rather than cutting out a coupon.
2. Experimental
2.1. Fractography
The surface was examined and documented for the two
pipe segments A1 and A2 in the as received state to document the location
and appearance of the SCC. Typical fracture surfaces were prepared for
exami- nation as follows. The pipe segments were sectioned into smaller
segments, these segments were broken open, and the fracture surfaces were
examined using optical and SEM. This also allowed the SCC to be examined
on polished sections. SEM examination was also carried out of the
comparison fracture surfaces from laboratory specimens soon after active
SCC in the laboratory.
2.2. Depth profiling
Depth profiling made use of the extreme surface
sensitivity of AES. That is, the AES technique provides information from a
very thin surface layer ~ 1 nm in thickness. The AES apparatus included a
sputter gun that allowed sputtering the surface over a controlled area.
That is, the material on the surface within a circle ~ 0.5 mm in diameter
was progressively removed, atom layer by atom layer at a controlled rate.
Interrupted sputtering combined with periodic analysis
allowed depth profiling in terms of a measurement of oxygen concentration
[O] and iron content [Fe] as a function of sputter time or depth. The
sputter time was converted to an equivalent depth using calibration
experiments to measure the sputter rate through Ta2O5
[1]. The depth profiles were interpreted to give a thickness for the
surface oxide layer.
2.3. Laboratory SCC samples
Fresh stress corrosion cracks were produced in the
laboratory using the LIST apparatus [2]. The LIST samples JQ1 and JQ2 were
machined from X65 pipelinesteel. One side had an ‘‘original’’ surface and
the other side was electropolished. The applied potential was -600 mV
selected according to the fast and slow potentiodynamic scans as shown in
Fig. 1. Two kinds of LIST test were performed as shown in Fig. 2. For
sample JQ1, the constant loading rate was 1.91x10-3 MPa/s,
until the stress exceeded the initiation stress as determined in reference
[3]. Thereafter the loading rate was 1.91x10-4 MPa/s until the
sample necked. After the sample necked, the stress was maintained for 3
additional days before the sample was pulled to fracture. Sample JQ2 was
loaded at a constant loading rate of 1.91x10-3 MPa/s until the
stress exceeded the initiation stress as determined in our prior work [3].
Thereafter the loading rate was 1.91x10-4 MPa/s for about 10 h
during the day and then kept constant overnight;then 1.91x10-4
MPa/s for about 10 h and then kept constant overnight. This process was
repeated four times until the sample necked during an overnight dwelling,
the stress was increased to fracture, the fracture surface was washed
using ethanol and distilled water, and then dried using a hair drier. The
fracture surface was observed using SEM.
|