Guides / Process
Reading a Broken Ceramic Part: Where the Crack Started and Why
September 23, 2026
A ceramic part does not deform before it fails, so there is no bent flange to tell you what happened. The information is in the broken surfaces instead, and it is usually enough to name the cause and stop it repeating.
Work through this before you reorder. Replacing the part without changing anything is the most expensive option available.
Step 1: find the origin
Fracture surfaces in a brittle material fan outward from the point where the crack started. Hold the pieces under a light at a low angle and look for the point all the ridges point back to. It is almost always at a surface, not in the middle.
Where the origin sits tells you most of the answer:
| Origin location | Usual cause |
|---|---|
| An edge or corner | A chip or a sharp undefined edge, often from handling or assembly |
| A hole, especially at the exit | Breakout chipping during machining, or a stress riser under load |
| A clamped face | Point loading from a bolt, a washer or a hard fixture face |
| A thin section | The design asked for something the material could not carry |
| A broad surface, no obvious feature | A material flaw, a thermal event, or grinding damage |
Step 2: run through the six usual causes
1. Edge chipping. The most common by far. An unchamfered edge, or a chip made during grinding, handling or assembly, becomes the crack origin. Fix it on the drawing with a defined chamfer, not with a note that says deburr. See sharp corners and edge chipping.
2. Point loading in the joint. A steel bolt tightened directly onto ceramic concentrates stress under the head. Spread it with a washer, a compliant gasket or a flange, and check the torque specification actually accounts for a brittle material.
3. Thermal shock. Fast heating or cooling puts the surface in tension against the bulk. Alumina and zirconia are both poor here. If the part cycles quickly, the fix is usually a different material, for example silicon nitride or cordierite, rather than a thicker section. See silicon carbide vs silicon nitride.
4. Expansion mismatch. A ceramic part clamped, shrink fitted or brazed into metal is loaded every time the assembly changes temperature. This is the classic cause of parts that pass testing and fail in the field months later. See metallising and brazing ceramic to metal.
5. Grinding damage. Aggressive grinding leaves microcracks below the surface that are deeper than the finish suggests. The part measures correctly and is weaker than the datasheet implies. Where a part is highly stressed, the finishing specification is a strength requirement.
6. A material flaw. Porosity, an inclusion or a firing crack. Real, but less common than the five above, and worth proving with a section or a CT scan before you blame the blank.
Step 3: check what changed
If the part ran for a year and then started failing, the material rarely changed on its own. Look for:
- A new blank lot or a new supplier for the blank
- A changed finishing step, for example grinding replacing lapping
- A new fixture, torque value or assembly sequence
- A process change upstream that altered temperature or chemistry
- Handling and packing changes, including a new freight route
This is why change control on a ceramic part is worth writing into the order. A silent grade or supplier substitution is a genuine failure mode.
Step 4: decide what to change
Most fixes are small and cheap compared with a second failure:
- Add or enlarge edge chamfers
- Add a radius in the internal corner where the crack started
- Spread the clamping load
- Increase a wall thickness that was below about 1 mm
- Specify a finer finish on the stressed surface
- Move to a tougher grade, for example ZTA or 3Y-TZP instead of alumina, covered in alumina or zirconia
Changing material is the last resort on the list, not the first. Geometry and handling explain more failures than grade choice does.
What to send us if you want help
Photographs of both fracture faces under raking light, the drawing, the grade and a sentence on what the part does. We read it, name the likely cause and match you with up to 3 shops for the replacement, with the change built into the quote. Nothing is charged to you.
Send the part details and we reply within 24 business hours.
Frequently asked questions
Can a cracked ceramic part be repaired?
No. There is no welding, brazing or straightening path that restores a structural ceramic. The part is replaced, which is why finding the cause before reordering matters.
Does a chip always lead to a crack?
Not always, but it is the most common starting point. A chip removes the intact surface and leaves a sharp notch in a material with no ability to yield, so the next load concentrates there.
How do I know if the failure was thermal or mechanical?
Thermal shock failures usually start at a surface that saw the fastest temperature change and often branch. Mechanical failures start at the highest stress point, typically an edge, a hole or a clamped face.
Need a quote for this part?
Send the drawing. We match you with up to 3 shops that grind this grade and reply within 24 business hours.
Get matched with a ceramic shop