Crazing is a glaze defect in which the fired glaze develops a web of hairline cracks, usually visible right out of the kiln but sometimes appearing weeks or years later. The underlying cause is a mismatch between how much the glaze and the clay body each expand and contract with temperature, not (as is commonly assumed) a matter of firing schedule or bad luck. It is the opposite problem to shivering, where a glaze is compressed too tightly and flakes off instead of cracking.
¶ The mechanism
Every fired ceramic material, glaze and clay body alike, has its own coefficient of thermal expansion (CTE)The fractional amount a material's length changes per degree of temperature change. Ceramic CTE values are small, typically expressed as a number times 10⁻⁶ per °C, but because fired ceramic is brittle, even a small mismatch between two bonded layers creates serious stress., describing how much it grows on heating and shrinks on cooling. Because the glaze layer is fused directly to the body, it has no choice but to follow the body's dimensional changes as the piece cools; it cannot expand or contract independently. Fired ceramic tolerates compression well but is much weaker under tension, so the direction of the mismatch determines which defect shows up: if the glaze wants to shrink more than the body allows, it ends up stretched (in tension) and eventually splits to relieve that stress. That splitting is crazing.
¶ Typical numbers
CTE values for ceramic materials are small enough that they are usually written as a coefficient times 10⁻⁶ per °C, and glaze-fit discussions often just quote that leading number. Measured examples from a working potter's own dilatometer testing give a concrete sense of scale:
| Material | CTE (×10⁻⁶ / K) |
|---|---|
| Tucker Smooth White stoneware | 6.64 |
| Tucker 6-50 porcelain | 7.71 |
| Mean of a wide sample of working glazes (Clayart archive) | 6.2 |
The rule of thumb long taught is that a stable glaze should sit roughly 1–10% below the clay body's expansion, just enough that the glaze ends up under mild, safe compression rather than tension. A review of successful glazes from the Clayart community archive, however, found expansions spread well outside that supposed 10% window, with functional pairings surviving at a much wider range than the rule implies. Treat the 1–10% figure as a starting estimate to test against, not a hard law.
¶ Adjusting the chemistry
Because thermal expansion is largely additive across a glaze's oxide makeup, shifting a recipe's chemistry, not just its firing schedule, is the actual lever for fixing a crazing problem:
The reverse toolkit applies to shivering, where the glaze needs raising rather than lowering; the underlying diagnostic is the same CTE-mismatch logic either way.
¶ Why it matters beyond looks
A crazed piece comes out of the kiln already pre-cracked, which gives any later stress a ready-made place to fail from. Measured strength reductions on freshly-fired crazed ware have been recorded in the 300–400% range compared to an equivalent well-fitted glaze, meaning glaze fit can matter more to a piece's durability than how hot it was fired. On functional ware, the network of cracks can also harbor moisture and bacteria over time, which is the standard reason crazed pieces are discouraged for food storage even when the crazing looks cosmetically minor.
¶ References
- Hansen, T. Glaze Crazing. Digitalfire. digitalfire.com & Understanding Thermal Expansion in Ceramic Glazes. Digitalfire. digitalfire.com
- Sankey, J. Pottery: Glaze Thermal Expansion (measured dilatometer data). johnsankey.ca/glazeexpansion.html
- Technofile: Glaze Fit, Ceramic Arts Network. ceramicartsnetwork.org
- Common Glaze Defects, Glazy. help.glazy.org/concepts/defects
- Chen, T., Gong, B., Tang, C. Origin and Evolution of Cracks in the Glaze Surface of a Ceramic during the Cooling Process. Materials 16(16):5508, 2023. doi.org/10.3390/ma16165508
- Photo by Andy Mabbett, via Wikimedia Commons, licensed CC BY-SA 4.0 (resized and re-compressed from the original).
.jpg){kind=link}