stress-corrosion cracking
Take a metal that shrugs off a mild corrosive on its own, and take a stress well below what would ever bend or break it — each harmless by itself. Put the two together in the wrong pairing and the metal can split open along a fine crack with almost no warning. That is stress-corrosion cracking (SCC): a brittle-looking crack that grows under the combined action of a steady tensile stress and a specific corrosive environment, in a susceptible alloy. It is one of the most feared failure modes precisely because the part looks fine and the loads look safe.
Three conditions must line up, and it takes all three: a tensile stress (from service loads or, insidiously, from leftover residual stress after welding or bending), a susceptible alloy, and a specific chemical environment that alloy is sensitive to. The pairings are oddly specific — austenitic stainless steel cracks in hot chlorides, brass cracks in ammonia (the historic 'season cracking' of cartridge cases), mild steel cracks in hot caustic ('caustic embrittlement' of old boilers), and aluminum alloys in certain chloride solutions. Corrosion opens a tiny pit or notch, the stress concentrates at its tip, that concentration drives corrosion and cracking a little deeper, and the crack ratchets forward, often branching and running between or through grains, with very little overall metal loss.
SCC is dangerous because the crack advances slowly and hidden, then the remaining section finally snaps fast, so failure feels sudden even though it was months or years in the making. Because all three factors are needed, removing any one defends against it: lower the tensile stress (design changes, or a stress-relief heat treatment or shot peening to erase residual tension), change the environment (remove chlorides, dose an inhibitor, drop the temperature), or switch to an alloy not susceptible to that specific medium. Note it lives at the border between corrosion and fracture — it needs the environment, so it belongs here, not with purely mechanical cracking.
Austenitic stainless steel is the textbook victim: an insulated hot pipe wrapped in wet chloride-containing lagging can crack right through in service, even though the pipe stress is modest, because trapped hot chloride under the insulation supplies the exact environment its residual welding stresses need to start chloride SCC.
Residual welding stress plus trapped chloride under insulation is a classic SCC recipe — no external overload required.
SCC is not ordinary overload fracture with a bit of rust: the stress alone would never break the part and the environment alone would barely corrode it, so a standard strength calculation that ignores the environment can call a design 'safe' that will crack.