Observations of corrosion pits and cracks in corrosion fatigue of high strength aluminum alloy by computed-tomography using synchrotron radiation

Measurements of the shape and dimensions of the pits and cracks formed in corrosion fatigue tests of a high-strength aluminium alloy, 7075-T651 were conducted by computed-tomography using ultra bright synchrotron radiation X-ray. Complex corrosion process could be observed and detail shape of corrosion pits could be identified. Sometimes, corrosion pit was formed under passive film, and those were not observed from the surface.


Introduction
The strength of metallic materials in corrosive environment is usually much lower than that in inert environment, and the stress enhances the corrosion damage because of the rupture of the passive film, which reduces the exposure of bare metal surface to the deleterious environment.In fatigue of metals in air, cracks are usually initiated from slip bands those are formed at the surface.In corrosive environment, the passive film is ruptured by the slip of grain, and then local cell is formed, where the bare metal face is anode and the passive film is cathode.As a result, a small hole, called the corrosion pit, is formed, and it gradually grows, finally cracks are initiated from the pits.Therefore, it is important to examine the formation and growth behavior of the corrosion pit to understand the corrosion fatigue damage.
Although the growth behavior of the corrosion pit in the corrosion fatigue process has been observed by optical microscopy, three dimensional shape of the pit could not be identified.Komai and his co-workers developed a stereo fractography technique by scanning electron microscopy (SEM) [1].In this method, however, the interruption of the fatigue test is unavoidable, and a sample has to be observed in vacuum.They also conducted in-situ observation of the growth behavior of corrosion pit by scanning atomic force microscopy (AFM) in the corrosive environment [2].Nakai and his co-workers conducted AFM observations of pits those were formed in corrosion fatigue [3,4].However, only the surfaces with roughness less than few μm can be observed by AFM, although the corrosion pit formed in corrosion fatigue grows up more than 10 μm in depth.Therefore, only the initial stage of corrosion fatigue process could be studied by AFM.
High-strength aluminium alloy is a very useful material for components where high specific strength is required.In corrosive environment, however, the fatigue strength of the alloy is usually very low because of the formation of the corrosion pits.The shape of the pits depended on the microstructure, and they were very complicated [5].Then it is usually difficult to determine the 14th International Conference on Experimental Mechanics stress concentration factor, local environment, and crack initiation site in the corrosion pit.The purpose of the present paper is to identify the shape of the corrosion pits three dimensionally and crack initiation site in a high-strength aluminium alloy under corrosion fatigue.Micro computedtomography using the ultra-bright synchrotron radiation X-ray (SR-μCT) was employed for the observation.The authors have been applied the SR-μCT to the 3D measurement of inclusions and cracks in steels, and a titanium alloy [6][7][8][9].

Experimental procedures
The materials employed for the present study was a high-strength aluminium alloy, 7075-T651.The chemical compositions and mechanical properties of the alloys are shown in Tables 1 and 2, respectively.The shapes of the alloys when they were received were plates with thickness of 11 mm and   14th International Conference on Experimental Mechanics SR-μCT imaging was carried out at beam line, BL19B2, of SPring-8, which is the world's large synchrotron radiation facility.X-ray energy was adjusted to 28 keV with silicon double-crystal monochromator.The distance between a bending magnet (X-ray source) and the specimen was about 100 m.X-ray area detector was set by 150 mm behind the sample.The projection image of penetrated X-ray was observed by an X-ray area detector.The detector was composed of a beam monitor and cooled CCD camera.Transmitted X-ray is converted to visible light through a thin phosphor screen and projected to the CCD camera by an optical relay-lens.The size of the pixel in the transmitted projection image was 1.37 μm for Material B or 0.70 μm for Material C. For 3D reconstruction of Material B, a set of 600 radiographs of a sample were recorded over 180 o rotations, where each rotation angle was 0.3 o .The angle was 0.5 o for 360 radiographs of Material C. Slice images were reconstructed from the series of projection images by filtered-back projection algorithm.

Artifact generated by corrosion chamber
For the CT imaging of aluminum alloy specimen, more than 1h was needed to obtain data for 3D reconstruction because it took 10 s of X-ray exposure for recording every radiograph.To avoid the progress of corrosion during the data acquisition, the observation was carried out without immersing  In principle, it is necessary that the overall observation area, including the sample and the corrosion chamber, should be set in the recording area of CCD camera for CT imaging.Since large recording area results in the loss of the resolution for the imaging, only specimen observation area was set in the recording area.At first, we expected that the effect of corrosion chamber may be negligible because the X-ray absorption by silicone rubber is much smaller than that by aluminum alloy, and we tried Type I corrosion chamber, whose width is bigger than the recording area The cross section image of CT with Type I and Type II chambers are shown in Figures 5 (a) and (b).For Type I chamber, artifact, which crosses the section of the sample, appears although the artifact exist only outside of the sample in Figure 5 (b) for Type II chamber.The difference between Types I and II chamber is that in X-ray penetration thickness of the chamber is almost constant for Type II chamber, and it is not for Type I chamber.The thickness of the silicon rubber changes discontinuously in Type I chamber as shown in Figure 5 (a).There still appear artifacts outside of the cross section of the sample as shown in Figure 5 (b) for Type II chamber although no artifact appears in Figure 5 (c), which was obtained without chamber.The elliptical shape of the cross section of the chamber may have brought the artifact because the penetration thickness of the silicone rubber is not exactly uniform.For Type III chamber, where both chamber and sample have circular cross section, no artifact is appeared as shown in Figure (d) despite only part of the chamber was in the record area.
In the cross section images, circumferential patterns, whose center agrees with the sample rotation center for CT, appeared.These patterns are also considered to be artifact.This kind of artifact must come from random noise of the measurement system, and could not be eliminated.14th International Conference on Experimental Mechanics

Comparison of CT images with SEM images
A SEM image of the surface of Material B in corrosion fatigue is shown in Figure 6.The surface is covered with a passive film, and the film is cracked and shows hexagonal pattern, where some part of the film is peeling off.An example of CT images of the same specimen are shown in Figure 7, where (a) and (b) is in a passive film, (c) is near the interface between passive film and base metal, and (d) is in base metal.Small cracks are also observed in the passive film layer observed by CT.It is important to note that the corrosion pits exist under the passive film, those were not found by the conventional microscopy.After 5×10 4 cycles from this observation, the passive film was peeled off and the pit could be observed by optical microscopy.In this alloy, successive formation and peeling off of the passive film was found.

Three-dimensional shape of corrosion pit
The CT image of corrosion pit was binarized and three dimensionally displayed in Figure 8.This pit was formed from an inclusion and the depth of the pit was 10μm.14th International Conference on Experimental Mechanics

Formation of corrosion pit
Detail observations of corrosion pits were conducted by circular cross sectional specimen with diameter of 1.4 mm (Material C).Owing to the slight difference between chemical composition of Materials B and C, the property of passive film is different, i.e., in Material C, hexagonally cracking and peeling off of the passive film were not observed.A corrosion pit formed under stress amplitude of 115MPa at number of cycles, N, of 4.9×10 5 is indicated in Figures 9 and 10, where Figure 9 shows the slice image, and Figure 10 is a three dimensional demonstration of the binarized image.
In Figure 9, the corrosion pit was extends to the longitudinal direction (rolling direction).Since the pit was covered with passive film on the surface, it may have not be observed from the surface observation by conventional microscopies, but surface at the pit was swelling up.Such swelling has been recognized in the AFM observation of the martensitic stainless steel [3,4].Inside of the material, this pit extended like root of plants.It suggests that anodic dissolution occurred along axially elongated inclusions or grain boundary.Another example of corrosion pit is shown in Figure 11.The shape of the corroded area is similar to that shown in Figures 9 and 10, and the swelling on the surface is more remarkable.These findings indicate that the corrosion speed was high where inclusions reached to the surface.

Growth of corrosion pit
Growth behavior of corrosion pit, shown in Figure 9, is displayed in Figures 12 and 13.The increase of corroded area and depth with number of cycles (or with time) could be measured from Figures 10, 12 and 13, and the shape of the corrosion pit shown in Figure 13 is similar to that observed by Wei and his co-workers [5].14th International Conference on Experimental Mechanics

Initiation of crack
Initiation process of cracks from corrosion pits shown in Figure 13 is indicated in Figure 14, where CT images of corrosion pit at the crack initiation site before and after the crack initiation is presented.From CT images of specimen surface (x-z plane), which is shown in Figure 14, indicate that crack was not formed from the largest corrosion pit at the surface.Otherwise, it was formed from large corroded area which was not necessarily deep.In the section image (y-z plane) which is perpendicular on the surface, as shown in Figure 14, the corroded region exists inside of the material ahead of the corrosion pit, and a crack initiated at the deepest point of the corroded area, whose depth was 40 μm, although the depth of corrosion pit was 12μm.14th International Conference on Experimental Mechanics

Conclusions
Measurements of the shape and dimensions of the pits and cracks formed in corrosion fatigue tests of a high-strength aluminium alloy, 7075-T651 were conducted by computed-tomography using ultra bright synchrotron radiation X-ray.The results obtained are as follows.
(1) Complicated corrosion phenomenon could be observed by computed tomography imaging using ultra-bright synchrotron radiation, and detail shape of corrosion pits can be identified.The effect of metal microstructure on the corrosion fatigue process can also be evaluated.
3.0 mm for Materials A and B, respectively, and it was round bar with diameter of 10 mm for Material C. The shapes and dimensions of the specimens are shown in Figure 1.A computer controlled electro-dynamic test machines were employed for the fatigue tests, where cyclic bending was applied to Materials A and B, and cyclic axial-force to Material C. Fatigue tests were conducted in 3.0% NaCl solution at a stress ratio, R, of -1, with loading frequency of 30 Hz for Materials A and C and 20 Hz for Material B. The corrosion chambers are shown in Figure 2. It was made of silicone rubber tube for Material B, and heat shrinkage tube for Material C. The corrosion fatigue tests for Material A were conducted in our pervious study to observe formation of pits by AFM, but it was not employed for the present CT observation.The S-N curves for Materials A and B are shown in Figure 4.The corrosion fatigue life of them is almost identical.

Fig. 1
Fig.1 Shape and dimensions of specimens (in mm, Arrows indicate rolling direction).
(a) Type I (b) Type II Fig. 3 Type III corrosion cell formaterial C. Fig. 2 Corrosion cell for Material B.

Fig. 5
Fig. 5 Effect of corrosion cell on CT imaging.
(a) In the circumferential section (parallel to the surface).(b) In the transverse section.
(a) In the circumferential section (parallel to the surface).(b)In the transverse section.

Table 2
Mechanical properties of materials.