Contact Us
Solve Your Materials Challenges
Find out how we can help
The total life time during fatigue generally can be divided into two regimes: one for crack initiation and another one for crack propagation. In the case of precipitation hardened alloys both regimes are influenced markedly by the microstructure of the material.
Important microstructural features are second phase particles such as large inclusions (> 5 µm), small dispersoids (0.1 to1 µm) and the strengthening precipitates (< 20 µm). The distribution of such particles can be influenced by manufacturing processes and heat treatments which may influence other microstructural parameters as grain size, degree of recrystallization and texture.
It is well known that grain size may influence mechanical properties of metals and alloys. When testing aluminum alloys it has been found that flow stress varies only slightly with grain size. Ductility, tensile strength and toughness, however, increase when grain size is reduced. The influence of grain size on fatigue properties has received less attention. Experiments have shown that fatigue life increases when grain size is reduced from 200 µm to 30 µm in an AlMgZn alloy in underaged and maximum hardness condition. The same has been found in AlCuMg alloys.
The automotive industry has during the last years made considerable efforts to increase fuel economy. One way has been to substitute to days materials, which normally are low alloyed carbon steels, by lighter materials, e.g. aluminum alloys. AlMgSi alloys (AA-6xxx series) are often chosen for these purposes because of good strength and ductility properties combined with high corrosion resistance and good hot and cold formability.
Crack initiation during fatigue is promoted, for example, by microstructures which tend to develop an inhomogeneous deformation which is usually more pronounced in coarse grained material. On the other hand, the crack propagation rate of such cracks along slip bands may slow down in such material because of a high reversibility of slip in front of the crack tip as proposed by Hornbogen and Zum Gahr. In microstructures with a more homogeneous slip distribution, cracks will mainly nucleate at large inclusions. The resistance to fatigue crack initiation may be better than it is for in homogeneously deforming microstructures, but the crack propagation resistance can be lower in this case.
Investigation of a pure and a commercial aluminum alloy containing 1 wt% Mg and 1 wt% Si have been selected. The commercial alloy contained also 0.8 wt% Mn and 0.4 wt% Fe.
All alloys mentioned here showed large inclusions of Mg2Si particles 5 to 20 µm in dimension which had an oval shape and were elongated in the extrusion direction. The commercial alloys also contained large inclusions of another intermetallic phase enriched in Fe, Si and Mn. This second phase was larger in size and more blocky shaped than the former. Dispersoids in the size range 0.1 to 1 µm were observed only in the commercial alloys. They contained Mn and differed in their dispersion.
Alloy CC showed a coarse dispersion with particle sizes of the order of 0.5 µm. These particles could not prevent recrystallization during extrusion. Therefore, large pancake shaped grains resulted with an average grain size of 400 µm. In contrast, alloy CF contained finely dispersed particles with an average size of 0.1 µm. The resulting grain structure was non-recrystallized with a 5 to 15 µm subgrain size. The absence of dispersoids in the pure alloy resulted in a globular grain structure with an average size of 90 µm.
Fatigue Crack Initiation. The number of cycles to crack initiation is defined as the number of cycles when a crack of 5 to 10 µm length could first be detected on the notch surface.
The largest cracks during the initiation stage are formed in the pure alloy. They nucleate either in persistent slip bands or at grain boundaries. The slip bands have the length of the grain diameter (90 µm) because there are no dispersoids hindering their development. The larger slip length and the coarse slip distribution result in high slip offsets on the surface leading to a fast crack nucleation. The nucleation at grain boundaries probably is caused by precipitation free zones (PSBs) adjacent to the boundary. A close inspection of intergranular fracture surfaces reveal fine dimples indicating strong local deformation in these zones. The third possible source for crack nucleation, namely the large Mg2Si inclusions, does not play a role in this alloy as well as in the other ones.
Fatigue Crack Propagation. The best overall fatigue crack growth resistance is observed for commercial alloy CC. For the pure alloy and for low the stress intensity amplitude ΔK values transgranular slip band cracking is mainly observed with coarse steps in the fracture surface. With increasing fracture surface examinations by scanning electron microscopy (SEM) showed that ΔK intergranular fracture becomes dominant. Very fine dimples on the intergranular fracture surface indicate plastic deformation in this region as already mentioned before. This change in the fracture mode can be explained by the varying size of the plastic zone in front of the crack tip.
The preceding results on fatigue crack initiation and propagation showed that for pure alloy the whole life time is determined only by the crack propagation period since the nucleation period is restricted to the first few cycles of the experiment. For the commercial alloy the total fatigue life is the same for its both versions commercial alloys. However, the fatigue initiation and propagation periods are quite different. For the coarse (CC) microstructure version about 60% of the life time is used up for crack initiation where as more than 95% is needed for the fine (CF) version microstructure.
This result may lead to the idea to improve the fatigue property of the commercial alloy C in such a way that the microstructure CF is used in areas where crack initiation is expected while the bulk of the material consists of microstructure CC leading to a higher resistance against crack propagation.
Fatigue is usually divided into four stages; crack initiation, stage I and stage II crack growth and rest fracture. In the tested specimens the rest fracture occurred in the same fibrous grain structure and covered the same area fraction. The rest fracture therefore did not seem to vary from specimens with fibrous grain structure to specimens with mixed grain structure.
As crack growth was not measured it is difficult to be certain whether the better fatigue properties observed for fibrous material is due to crack initiation or crack growth. But fractographic investigations showed that cracks had started in the layer of coarse, recrystallized grains in specimens having a mixed grain structure. And as the layer of recrystallized coarse grains are thin, < 400µm, the main length of stage II crack growth must therefore take place in material having a fibrous grain structure, also in specimens with a mixed grain structure. The conditions for stage II crack growth should therefore be much: the same in materials of both grain structures. The observed difference in number of cycles to failure is therefore most likely due to easier crack initiation and stage I crack growth in specimens having a thin surface layer of coarse recrystallized grains.
Other reasons for easier crack initiation in the coarse, recrystallized grains than in fibrous grain structure may be due to the morphologie of hardening particles in age hardening aluminum alloys. In these alloys a precipitate free zone (PFZ) is present adjacent to grain boundaries. This zone has therefore a lower strength than the grain interior and deforms at a lower stress.
When using age hardening aluminum alloys for purposes where fatigue and corrosion properties are of importance, the grain structure of the material may therefore be of importance even though no influence is found when testing strength, ductility and toughness.
Cyclic stress-strain data, and crack resistance curves under monotonic loading, have also been recorded. The fracture surfaces were examined using a TEMSCAN electron microscope in both scanning (SEM) and transmission (TEM) modes. This technique was employed to investigate both the crack tip deformation structure, and crystallographic orientation of fatigue cracks. X-ray microanalysis was used to determine the composition of particles on the fracture surface.
A great deal of scatter is apparent, and information from small amounts of data, particularly close to the resolution limit, can be misleading. Consequently, includes some data from other sources to indicate broad trends. The following observations are emphasized:
SEM and TEM observations give evidence of two morphologically distinguishable modes of crack extension. In a laboratory air environment the first results in a fracture surface which is crystallographic in nature, and which exhibits fatigue striations in many areas. At low ΔK, this mode of growth dominates resulting in a fracture surface composed almost entirely of approximately crystallographic facets. The second mode of extension is one of micro-void coalescence resulting in a typical dimpled fracture surface, regions of such fracture often being associated with large inclusions.
It is proposed that crack growth under fatigue conditions, in high strength aluminium alloys, is the result of three micromechanisms acting together:
If there is an alternative mechanism available for crack extension, the crack may grow by that process until its tip is close enough to a void to satisfy the ligament instability condition, and only then extend a short distance by the formation of a dimple. Under these conditions a new probability density function is defined.
At these lower stress intensity factors, the major contribution to crack growth is clearly from a mechanism resulting in crystallographic fracture, and, in many cases, fatigue striations.