How Each Common Metal Tells You It Is Failing

Why this matters

Every metal in a building fails in its own handwriting. Steel goes orange and flaky and pushes joints apart as it grows. Cast iron does the opposite and fails while still looking perfect. Brass changes color from yellow to pink before it lets go. If you can read four or five of these signatures you stop guessing at remaining life, and more usefully, you get several independent witnesses to the same room: when the aluminum, the galvanized, and the stainless in one space all report the same aggressor, you have named the cause rather than the symptom, and the fix stops being a parts list.

Before you read corrosion inside electrical equipment

Corroded terminals and green-crusted lugs are common findings and they sit inside energized enclosures. De-energize the circuit, apply your lock and tag, and then prove the conductors dead before any part of you or your light enters the enclosure. For work on a panel, a branch circuit, or energized conductors, the controlling standard is 29 CFR 1910.333(b)(2), not the general energy-control standard. The proving sequence is live-dead-live: test your meter on a known live source, test the conductors you are about to touch, then test the known live source again to confirm the meter still works (NFPA 70E-2021, 120.5). A meter that failed silently between the first and second reading is the failure mode that sequence exists to catch.

Where the stored energy is mechanical or fluid rather than electrical, a spring, an accumulator, or a pressurized vessel, the isolation and verification requirement is 29 CFR 1910.147 instead.

Carbon and mild steel: it grows

Steel rusts to a flaky orange-brown product that occupies several times the volume of the metal it consumed. That expansion is the part of the story people miss. Rust in a lap joint, under a bracket, or between a steel plate and a bolted flange acts like a slow jack: it splits welds, bows plates, cracks masonry that steel is embedded in, and pops coatings from underneath.

• Loose orange flake with sound metal beneath: surface attack, mostly cosmetic, and the useful action is finding the water source.
• Layered scale you can lift off in plates: significant section loss. Estimate remaining thickness before you trust the part.
• A coating blistering or a joint opening with no obvious external rust: rust is working underneath. The visible damage is the last stage, not the first.

Galvanized steel: the coating is a clock

Zinc protects steel two ways, as a barrier and sacrificially, which means it corrodes on purpose. Read it in that order:

• White powdery product ("white rust"), often after storage in a wet stack: zinc corroding. The coating is being consumed but is still doing its job.
• Dull grey, no white, no red: normal weathered zinc. Nothing to do.
• Red rust in patches: the zinc is locally gone and the steel underneath is now unprotected. This is the moment the clock starts, and its position tells you where the water is.
• Red rust at cut edges only: normal and slow at first, because zinc protects a small distance of bare steel at an edge sacrificially. Widespread red away from edges is not that.

Zinc consumption rate depends entirely on the environment, so do not carry a service-life number between sites. What travels is the sequence: white, then grey, then red, then section loss.

Stainless steel: staining and pitting are two different findings

Stainless resists corrosion because of a passive oxide film. Two things break it and they look alike at ten feet.

• Rust-colored staining that wipes or polishes off, with sound metal underneath, is very often free iron contamination rather than the stainless corroding: carbon steel particles from a grinder, a steel brush, or steel swarf, rusting on the surface. The fix is cleaning and passivation, and the lesson is to keep carbon steel tooling away from stainless.
• Discrete pits, often under a deposit or inside a crevice, are the stainless itself failing, and chlorides are almost always the driver. A little rust bleeding from a single point with clean metal all around it is far more serious than a broad tea-colored stain.
• Fine branching cracks, sometimes with almost no corrosion product, are chloride stress corrosion cracking, which needs tensile stress, chlorides, and heat together. It is commonly described as becoming a practical concern above roughly 60 C (140 F) for the standard austenitic grades, and it produces sudden through-wall failure in a part that looks nearly clean. Treat any cracking in hot chloride-bearing service as a replacement decision, not a cleaning one.

Aluminum: white powder, and it hates company

Aluminum forms its own protective oxide and then fails in three characteristic ways:

• White to grey powdery product, localized: oxide breakdown, usually where a deposit held moisture.
• Pits, sometimes deep, with the surrounding surface clean: the normal aluminum failure. Depth matters far more than area, and pit depth is measured and rated under ASTM G46 rather than judged by eye.
• Thread-like worm tracks under a coating (filiform corrosion): moisture creeping under paint or anodizing from a scratch or edge.

Aluminum is also the metal most often destroyed by its neighbors. Against copper or a more noble alloy with moisture present, it is the one that dissolves. In contact with fresh concrete, mortar, or other strongly alkaline material, it is attacked directly.

Copper and brass: color is the diagnosis

• Copper with green or blue-green crust (verdigris) on the outside: external moisture, and often a joint that has been weeping long enough to leave a trail.
• Copper with pinhole leaks and horseshoe or U-shaped pits on the inside, pointing downstream: erosion-corrosion from excessive water velocity, usually just past an elbow, a partly closed valve, or a burr left at a tube cut. The shape of the pit is the direction of flow.
• Copper that leaks at a pinhole with no external warning and shows a network of fine tunnels in a cut section: formicary or ant-nest corrosion, driven by trace organic acids in the surrounding air rather than by the water inside. The give-away is that the outside surface looks fine right up to the leak.
• Brass that has turned from yellow to pink or copper-colored, feels soft, and is porous: dezincification. The zinc has leached out and left a spongy copper skeleton that keeps the part's shape and none of its strength. Often accompanied by white or blue-white deposits at the leak. This is a water-chemistry finding, and the countermeasure is a dezincification-resistant brass or a bronze, not a tighter fitting.

Cast iron: the failure that looks like nothing

Grey cast iron in wet service undergoes graphitic corrosion. The iron dissolves and leaves behind the graphite network, so the part keeps its exact shape, its dimensions, and often its paint, while losing nearly all of its strength.

The field test is simple and it is the only reliable one: press a knife point or a screwdriver into the suspect area. Sound cast iron resists. Graphitized cast iron takes the blade like hard soap and comes away as a dark grey, sooty material. Any pressure-containing cast iron that fails this test is a replacement, immediately, because its failure mode is sudden and full-bore rather than a weep.

Zinc die-cast parts have their own quiet version: internal corrosion makes the part swell and crack with almost no external product, so a housing or a handle that has grown enough to bind, or cracked with no impact, is telling you the alloy is going rather than that someone forced it.

Reading a room across four metals

An equipment room contains a mixed set of hardware and a service call for one corroded panel. The tech de-energizes and proves dead before opening it, finds green crust on two terminals, and instead of just remaking them, walks the room.

What the metals say. The galvanized housings show red rust in a continuous band from the floor to roughly 3 ft up, and dull grey zinc above it. The stainless fasteners show discrete pits with clean metal between them, not general staining. An aluminum bracket at floor level is pitted through in two places while an identical bracket at ceiling height is clean. Copper tube shows green external crust at the wall penetration where it is lowest.

What that combination rules out. A leak overhead would put the worst damage under it and would run downward in streaks, not in a level band all the way around the room. General humidity would attack the aluminum and the galvanized fairly evenly with height and would not put discrete pits in stainless, which shrugs off plain moisture. Something with a strong height gradient and a chloride signature is doing this.

What was there. Chlorine-bearing cleaning and treatment chemicals stored open on the floor, in a room with no dedicated exhaust. The vapor is heavier than air, so it pools low, which produces exactly the level band, and chloride is precisely the agent that pits stainless and aluminum while accelerating everything else.

The scale of it. Of the seven pieces of equipment in the room, 5 sit with some part of their enclosure inside that lower band, and all 5 show band corrosion. The 2 that are wall-mounted above it do not. That is a clean split, and it is the single strongest piece of evidence, stronger than any individual metal's appearance.

What it cost to miss. The shop had remade terminals in that room three times over about two years, roughly 1.5 hours each, so about 4.5 hours of repeat work on a symptom. Moving the chemical storage out of the room and adding exhaust is a one-time action. The corrosion already present does not reverse, so the enclosures on the worst two units still need replacement, but the recurrence stops.

What would change the reading. If the band had been at the ceiling rather than the floor, the source would be something rising, such as combustion products or a vapor lighter than air, and the storage would be irrelevant. If the stainless had shown broad tea-colored staining rather than discrete pits, the finding would more likely be iron contamination from installation tooling, which does not implicate the room at all. And if only one metal had shown anything, there would be no room-level conclusion to draw from a single witness.

Testing your read before you commit to it

• Find the control. Every conclusion above depended on an identical part in a different position. Before you name a room-level cause, locate a piece of the same metal that should be clean under your theory, and confirm it is.
• Probe rather than photograph on cast iron and on anything you plan to keep in pressure service. Appearance is the one thing graphitic corrosion does not affect.
• Measure remaining thickness where the call is repair or replace. An ultrasonic thickness gauge reads through paint and rust on steel and turns an argument into a number.
• Scrape before you conclude on stainless. Staining that comes off with a nylon pad and a cleaner is contamination. Staining that keeps bleeding from a fixed point is a pit.
• Photograph the height and the orientation, not just the damage. A close-up of corrosion proves the corrosion. Only the wide shot proves the pattern, and the pattern is what names the cause.

References

• 29 CFR 1910.333(b)(2), safety-related work practices for electrical work, for the de-energize, lock, tag and verify sequence before opening an energized enclosure
• NFPA 70E-2021, 120.5, for the live-dead-live instrument proving sequence
• 29 CFR 1910.147, control of hazardous energy, where the stored energy is mechanical or fluid rather than electrical
• ASTM G46, examination and evaluation of pitting corrosion, for pit depth measurement and rating
• See related: Reading Rust and Corrosion Patterns; The Galvanic Series as a Field Tool; Pitting, Crevice and Uniform Corrosion