In many failure investigations, interpretation of the features on the fracture surface (or surfaces) can be extremely useful in determining the mode of failure. One of the most interesting and revealing features is that of micro-void coalescence (MVC). Micro-void coalescence is a high energy fracture mechanism in metals, which is very common and readily interpreted. The process occurs effectively in three stages: i) micro-void nucleation, ii) void growth, and iii) coalescence.
In a material under increasing (tensile) stress, there is initially the nucleation of microscopic voids around typical strengthening alloying elemental particles (such as manganese) in an otherwise more ductile matrix of the steel. As the load increases, or at the increasing stress field at the tip of a crack, these micro-voids grow, eventually linking up to form a growing crack front. At the stage where the growing micro-voids link up they are characterized by localised micro plasticity and regions of localised necking between adjacent micro-voids. In pure tension, the appearance on a micro-structural scale is typically of an approximately circular region (or cusp), surrounded by a lip of plastically deformed material linking to the next void, (as shown in the SEM image). Under pure tensile loading stress conditions, this is typically manifested by nearly circular regions. With care one can sometimes see the central stiffening or alloying particle at the approximate centre of such micro-voids. The micro-void size can vary, in proportion to the amount of the alloying constituent, leaving a range of approximately circular tensile fracture surfaces with different sizes. The analogy has been made that this is rather like toffee/nougat being pulled apart, with voids initiating at the ‘pieces of nut’ with necking occurring around the edges.
This phenomenon is well understood and documented, and is remarkably helpful in interpreting fracture surfaces. On the one hand, MVC clearly indicates ductile behaviour that is easily recognised, and in the case shown in the image, indicative of tensile loading. More importantly, however, is that MVC behaviour under shear and tearing stress situations is also particularly useful in manifesting stress behaviour. As mentioned above, nominally circular micro-voids are indicative of pure tensile stresses, but if the stresses were to be of a more shear nature, when the voids are pulled apart and link up, the ductile edge ligaments are pulled apart in a shear manner. This results in the circular areas of plasticity being distorted in shear and is manifested at the edges as a ‘C’ shape rather than the circular shape of tensile MVC. In such shear loading cases the series of such ‘C shapes’ all point in the same direction on the one surface, and in the opposite direction on the mating (opposite) surface, as is clearly apparent from the shear stress loading. Thus one has the two C shaped MVC pointing towards each other (open ends of the C shape facing each other on opposing mating fracture surfaces. Such surface appearance is very useful in interpreting that the final nature of fracture (in this case) was one of pure shear.
Similarly if the MVC ‘C shapes’ point in the same direction on opposing surfaces, then the stress case is clearly one of ‘tearing’ or ‘bending’, not shear, with the open ends of the capital C shape pointing toward the more closed end of such a tearing crack. Again this feature of microstructural behaviour is very useful in interpreting fracture mechanisms in failed structures, and can indeed be used to ascertain crack propagation direction in ductile tearing situations.
This latter case of tearing was particularly useful in a case involving the crash of a small (six seater, twin engined) light aircraft. The plane crashed in mountainous terrain and the pilot and all the passengers were killed. The subsequent investigation indicated that one wing had ‘torn off’ and it was initially thought that the wing had torn off on impact with the mountain-side, with the main crack propagating from front to back, and questions were raised about pilot error. More careful study of the fracture surface, however, revealed that the wing had indeed torn off, but in a ’forward’ direction (from a fatigue crack in a support spar). The main crack propagated from the back of the wing (trailing edge) towards the leading edge, as a result of which the wing (dragged by the engine) rotated forwards. The failure was clearly not due to pilot error at all, but simply (and sadly) a fatigue crack reaching its critical crack length, for catastrophic failure.