subcritical crack growth
A ceramic that survives a load today can quietly break under that very same load weeks, months, or years later, with nothing new having happened to it. This delayed, treacherous failure is subcritical crack growth. A glass shelf holds its books for years and then one morning is found cracked in two; a load that was safe has, over time, become fatal. The reason is that even below the stress needed for instant fracture, a crack does not always stay still. It can creep forward, slowly, until it finally reaches the critical size and the part fails suddenly.
The engine of this slow growth is usually chemistry, not mechanics. Water molecules from the air, or body fluids, or fuel, reach the highly stressed bonds right at the crack tip and react with them, breaking them one at a time in a process of stress-assisted corrosion. Each broken bond lets the crack inch forward, exposing fresh bonds for the next molecules to attack. The crack velocity depends steeply on the stress intensity, following roughly v = A times K^n, where n is the slow-crack-growth exponent; a large n, often between 15 and 50 for ceramics and glasses, means the material is more resistant, its cracks barely creeping until K is very close to K_IC. In ceramics this is called static fatigue, because failure comes from holding a constant load over time, which is quite different from the cyclic, repeated-loading fatigue of metals.
Subcritical crack growth matters because it means a one-time strength test cannot guarantee long-term safety; a part must be designed to a low enough working stress that its worst flaw will not creep to critical size within the required service life, and the environment (humidity, steam, body fluid) is a first-order variable in that lifetime. The steep dependence on stress, captured by n, is exactly what makes lifetime prediction possible and what underpins proof testing. An honest warning against a common confusion: static fatigue in ceramics is chemistry at the crack tip, not the cyclic plasticity that fatigues metals, so a ceramic can hold a steady load for years and then break with no external event to blame, which is precisely why it is so dangerous.
Optical glass fibre is proof-tested and used well below its instant strength precisely because of static fatigue: humidity slowly grows the tiny surface flaws over the fibre's decades of service. Engineers use the measured crack-velocity exponent n to predict how many years a fibre at a given strain will last before a crack reaches critical size.
Moisture slowly corrodes the stressed crack tip, so a load safe at first can grow a flaw to failure years later.
Static fatigue in ceramics is stress-assisted corrosion at the crack tip, not the cyclic-plasticity fatigue of metals. It means a ceramic can hold a load for years and then break with no new event, so a single strength test never proves long-term safety; you need the crack-velocity exponent n and a lifetime model.