Corrosion Failure Modes

Uniform corrosion is the electrochemical dissolution of metal more or less evenly across the exposed surface. It is the least deceptive mode: you can measure the section loss with a calliper or an ultrasonic gauge, you can estimate the corrosion rate in millimetres per year, and you can predict residual life. Failures attributed purely to uniform corrosion are usually management failures , the calculated service life was exceeded, or protective coatings broke down and were not repaired.

Galvanic corrosion introduces a spatial twist. When two metals with different positions on the galvanic series are in electrical contact in an electrolyte , seawater, condensed moisture, even high-humidity air , a current flows. The more active metal (the anode) corrodes preferentially. The practical importance is the area ratio: a large cathode coupled to a small anode drives severe, concentrated attack on the anodic metal. A bolt of a more active alloy in a large stainless plate can corrode through in a fraction of the time the uniform rate would predict.

Pitting is dangerous precisely because it is invisible from the outside for most of its life. The passive film on stainless steel or aluminium breaks down at a point , triggered by a chloride ion breaching the oxide, a surface inclusion, or a surface defect. Once a pit forms, its interior chemistry becomes self-sustaining: oxygen is depleted inside the pit while the cathodic reaction runs on the large surrounding surface, the pit environment acidifies, and local chloride concentration rises. The pit grows inward far faster than the general surface loss would suggest.

Crevice corrosion follows the same chemistry but is driven by geometry. Any tight gap , under a gasket, between overlapping plates, inside a threaded connection , restricts oxygen replenishment. The differential aeration between the crevice interior and the open surface establishes the same driven micro-cell. Stainless steels that are immune to pitting in open seawater can suffer severe crevice attack at flanges.

In a failure investigation, pits matter as fatigue crack initiators even when they have not themselves penetrated the section. A pit concentrates stress by roughly a factor of three at its base, enough to start a fatigue crack in a component that would otherwise have survived indefinitely. Finding a pit at the origin of a fatigue fracture is one of the cleaner causal chains in metallurgical failure analysis.

Stress-corrosion cracking occurs when three things coincide: a susceptible material, a specific corrosive environment for that alloy system, and a sustained tensile stress above a threshold. Remove any one leg and SCC stops. That is why the forensic task is to prove all three were present and acting together, not just one.

• Material susceptibility: high-strength steels in high-yield-strength conditions, austenitic stainless in chloride environments, brass in ammonia, aluminium alloys in sodium chloride. The environment-material pairing is specific.
• Stress source: not just applied service load. Residual stresses from welding, heat treatment, or manufacturing are often the hidden tensile stress that drives SCC in a component that carries little service load.
• Crack morphology: intergranular (crack follows grain boundaries) or transgranular (cuts across grains, often with a branching, feathery appearance under the SEM). The path depends on the alloy system and the specific environment.

The Silver Bridge at Point Pleasant, West Virginia, was an eyebar-chain suspension bridge built in 1928. Its design was unusual: instead of wire cables with many parallel load paths, the suspension system used chains of individual eyebars, each link carrying the full load. There was no redundancy. A single failed link brought everything down.

The National Bureau of Standards investigation, published in 1971, identified the failure in eyebar C13N. A stress-corrosion crack had grown from a corrosion pit on the inside of the eye head, where a finger of corrosive solution could penetrate the tight eyebar-pin contact. The steel was a heat-treated high-strength carbon steel with a yield strength of approximately 75,000 psi, well above mild structural steels of the era, and was susceptible to SCC in corrosive environments. The residual stress at the eye head, combined with the sustained bridge load, provided the tensile component. The Ohio Valley atmosphere , humid, mildly acidic from industrial pollution , was the corrosive environment.

The recovered fracture surface showed an approximately 2.5 mm deep SCC crack front, followed by a brittle fast-fracture zone that covered the rest of the section. There was no evidence of prior visual cracking or significant plastic deformation , the fracture was sudden once the crack reached the critical size for K_I to equal K_IC of the material. The investigation concluded that the bridge had never been inspected for internal crack development at the eye heads, a geometrically inaccessible location. The collapse led directly to federal bridge inspection standards in the US and elevated SCC to a first-order concern in high-strength steel infrastructure.

Corrosion fatigue is distinct from SCC in one important way: it does not require a specific environment-material pairing. Any corrosive medium that degrades the passive film, lowers the surface energy for crack initiation, or anodically dissolves crack-tip material can accelerate fatigue. The S-N curve shifts downward , meaning failure at stress amplitudes that would be safe in air. More significantly, the fatigue limit (the stress below which failure does not occur in air) effectively disappears in a corrosive medium: cracks can initiate and grow at arbitrarily low cyclic stress if the environment is aggressive enough and time is long enough.

Selective-phase corrosion attacks the microstructure itself rather than the surface uniformly. Dezincification removes zinc from brass, leaving a porous copper skeleton that looks intact but has lost most of its strength and ductility. Graphitisation of grey cast iron dissolves the iron matrix, leaving a graphite network. In both cases, cross-section measurements can be misleading: the section dimensions are unchanged but the material that remains is structurally compromised. Cutting a section and doing a microhardness traverse or an energy-dispersive X-ray analysis quickly reveals the depleted zone.

• Corrosion fatigue indicators: beach marks and striations visible as in air fatigue, but with rougher, more corroded crack faces, often with secondary cracking parallel to the main crack. Crack initiates at a surface pit rather than at a stress concentration alone.
• Dezincification indicator: pink or copper-coloured zone on a yellow-brass fitting, soft to a hardness probe, porous at 100x magnification.
• Graphitisation indicator: grey cast iron pipe with intact outer dimensions but zero sound velocity on ultrasonic testing; cross-section shows black graphite network without iron matrix.

Intergranular corrosion (IGC) preferentially attacks grain boundary regions rather than the grain interior. In austenitic stainless steels, the classical trigger is sensitisation: heating in the range 425 to 815 degrees Celsius , the welding heat-affected zone temperatures , causes chromium carbide to precipitate at grain boundaries, depleting the adjacent metal of chromium below the 12% threshold needed to maintain passivity. The boundaries become active relative to the grain interiors, and selective attack follows.

In a failure context, weld heat-affected zones are the first place to look for sensitisation in austenitic stainless. The Strauss test (ASTM A262 Practice E, boiling copper sulfate/sulfuric acid) and the Huey test (ASTM A262 Practice C, boiling nitric acid) are the laboratory evaluation methods. If the failure is in a sensitised zone, the next question is whether the material was specified correctly (low-carbon L-grade or stabilised Ti/Nb grade would have been resistant), whether post-weld solution annealing was carried out and verified, or whether there was a process upset that unexpectedly held the component in the sensitisation range.