What failure analysis is for
When a hard part breaks in service, the useful question is rarely “did it break?” but “why?” Was the material faulty, was the part damaged during handling, was the design stressing it more than expected, or was the environment harsher than planned? Failure analysis is the structured investigation that tries to answer that. The aim is to find the root cause so that the problem can be fixed, not simply to assign blame.
Brittle fracture leaves a readable record
Hard, brittle materials such as technical ceramics, cemented carbide and glass-like substances do not bend much before they break. A crack starts at one point, usually a flaw, and then spreads at high speed. The surface it leaves behind carries a record of that journey, and the study of those surfaces is called fractography.
The key idea is that fracture starts at a single origin, often much smaller than the part. The origin might be a pore, an inclusion, a machining scratch, a contact damage mark, or a grain that is unusually large. Because brittle strength is controlled by the worst flaw, the origin tells investigators a great deal about whether the problem lay in processing, finishing or use.
Markings on a fracture surface
Around the origin, brittle fracture surfaces in glasses and fine-grained ceramics commonly show a characteristic sequence:
- Mirror: a smooth, flat region surrounding the origin, formed while the crack was still moving relatively slowly.
- Mist: a slightly rougher zone just beyond the mirror as the crack accelerates.
- Hackle: coarse lines radiating outward as the crack speeds up and branches.
Hackle lines tend to point back toward the origin, like arrows. Larger mirrors generally correspond to lower breaking stress, so their size gives a rough indication of how heavily the part was loaded. Coarser-grained ceramics often show these features less clearly, and the fracture path may run along grain boundaries instead.
A typical investigation
The process usually follows a sequence that protects the evidence.
- Gather background. Investigators ask how the part was made, how it was installed, what loads, temperatures and fluids it saw, and how long it lasted. A part that failed on first use points to different causes than one that failed after thousands of cycles.
- Document before touching. Photographs of the pieces in their as-found position come first. Fracture surfaces are kept clean and are not rubbed back together, since contact can damage fine markings.
- Visual and low-magnification examination. A stereo microscope reveals the overall crack pattern, which piece broke first, and where the main branches run.
- Detailed imaging. Scanning electron microscopy images the origin, and elemental analysis can identify an inclusion or contamination.
- Supporting tests. Hardness, microstructure and chemical checks compare the failed part with what the specification called for. Sometimes companion parts from the same batch are tested.
- Synthesis. The investigator combines the evidence into a most likely sequence of events.
Common causes found
In hard materials, several recurring themes appear. Processing flaws such as pores or agglomerates left from the sintering process can act as origins. Surface damage from grinding or handling can create cracks that are not obvious by eye. Thermal shock, where a sudden temperature change creates internal stress, produces its own crack patterns. Mismatched thermal expansion in an assembly can overload a ceramic held rigidly by a metal part. Contact loading, impact and misalignment cause chipping at edges.
For layered or fibrous materials, the picture differs. Composites often fail by delamination, fiber breakage or matrix cracking, and the specific mix reflects the loading. The typical composite failure modes are well documented, and investigators use them as a reference for what they see.
For coated parts, failure may involve the film peeling away, which points toward coating adhesion or residual stress rather than the substrate itself.
Limits and honesty
Failure analysis does not always give a single clean answer. Evidence can be destroyed in the failure itself, and several contributing factors may have acted together. A careful report states what the evidence supports, what it merely suggests, and what remains unknown. It also notes whether other parts of the same type might share the problem, which is often the most valuable result of the whole exercise.