In tensile loading, crack-opening stresses create cracks that are perpendicular to the loading direction. In Figure 1, we see a crack that runs perpendicular to the length of the cylinder, or to the axis of the cylinder. This was a fatigue crack, and fatigue cracks are inherently brittle on the macro (visible) scale. Most people who know even a little bit about fracture analysis would recognize this as a brittle crack.
Figure 2 shows the threaded portion of a bolt that has obviously behaved in a very ductile manner on the macro scale. Note the readily visible reduction in cross-sectional area. The initiation area of the crack is perpendicular to the axis of the part, however, so we would consider that portion of the crack itself to be brittle on the macro scale. Once the crack has modified the stress state of the bolt sufficiently, we see a change of crack plane to one that is 45 degrees with respect to the loading and part axis. This portion of the crack is macro-scale ductile. A large fraction of fracture surfaces show ductile cracking in some portion of the crack and brittle in another.
Figure 3 shows a broken compact tension specimen, which also has a mixed-mode crack surface at the macro scale. Note that the area in the pink oval is the actual crack area. The lower left and far right very flat portions of the specimen extending from the fracture surface areas are machined flat in order to create the notch that is a part of the test procedure, intended to guarantee macro-scale brittle cracking.
In fact, the classical “ductile” cup and cone fracture that tensile testing routinely produces in properly heat-treated moderate hardness round-steel test coupons, technically speaking, is a mixed-mode fracture. The slanted walls of such a fracture are indeed macro-scale ductile. However, the flat bottom center area would be considered macro-scale brittle. What? This does not make sense! It goes against everything most people learn about ductile and brittle cracks. If you learn to think this way, however, you can much more reliably learn to distinguish ductile and brittle cracks in a manner that allows you the heat treater to defend yourself when someone tells you that you made their parts brittle.
Remember the point of macro-fractography is to identify the loading geometry so you can tell if the part was loaded the way it was supposed to have been loaded. If you look at the micro scale on the flat bottom of a macro-scale brittle cup-and-cone fracture, you will most likely see microvoids or “ductile dimples.” Just remember, the macro-scale information is about a totally different set of data than the microscale. The micro-scale data (microvoids) tells you how the material responded to the loads. The macro scale tells you the orientation or direction in which the loads were applied to the part.
Figure 4 shows a flat strip that was prepared as a tensile-test coupon. The material is thin and shows a classical slant oriented macro scale ductile fracture.
Axial-compression fracture photos are harder to find. Figure 5 shows a view of a part that has cracked in a macro-scale brittle manner due to internal pressurization, which stretches the part in the circumferential direction. Crack opening takes place in the same direction as the load is applied! Crack opening or macro-scale brittle cracks in compressive loading are parallel to the applied load direction, not transverse.
Figure 6 shows an actual macro-scale brittle crack feature on a tubular part that was compressed along its axis. A through-wall split in the axial direction, parallel to the loading direction, happened when the material bulged in a ductile manner! In the book Understanding How Components Fail by Don Wulpi (ASM International publication), there is a good photo of a little test coupon that was case hardened to make it behave in a brittle manner on the outer surface. The core was left soft. Four of the six faces of the cube were machined off and the load was applied to squeeze the two hard surfaces toward each other. The soft core bulges out due to macro-scale ductile behavior. If the loading continues, shear cracks at 45 degrees to the load can occur. The hard faces, however, show longitudinal cracks perpendicular to the direction of loading.