It is chalky, white and slightly raised, and it turns up in patches rather than evenly. You wipe it off and it returns. Underneath, the surface is rough and the colour has gone. On an anodised part the bloom often starts at a scratch, a cut edge or a machined face, where the anodic layer is thin or missing.
On a painted or powder coated part the same corrosion runs under the film and lifts it, and the white powder shows at an edge or through a blister. In both cases you are looking at aluminium being consumed rather than at dirt, and that distinction matters because the two need opposite responses.
Aluminium protects itself with a thin oxide that forms instantly in air, and that oxide is a good barrier while it is intact. The difficulty is that it occupies considerably more volume than the metal it came from, so when corrosion is confined beneath a film the product has nowhere to go and pushes the film off from below.
That is the mechanism behind the bloom. Once a breach lets moisture and chlorides reach bare metal, corrosion proceeds in a cell shielded from drying, the white product lifts the coating, more moisture gets in, and the cell widens. It is self-sustaining in a way that surface dirt never is.
An anodic layer is brittle. A scratch through it exposes bare metal, and a cut edge exposes the substrate along its length because the layer does not wrap around the cut. Threads and drilled holes are the same story: the film is thin on a thread crest and often absent inside the hole.
Welding is the largest breach. Heat destroys the anodic layer beside a weld, so a fabricated aluminium part carries a band of unprotected metal either side of every seam unless protection is restored after welding. On a coated part the same applies to any area heated, ground back or masked during finishing.
Residue wipes off and does not return. It is dried detergent, hard water or polish, and the metal underneath is smooth. Bloom wipes off, returns, and leaves a pitted or chalky surface behind. That one test separates the cosmetic problem from the structural one, and it needs no more than a wet cloth.
Filiform corrosion is finer and more directional: threads running under the coating, often from an edge, with a white head at the advancing end. Galvanic corrosion sits in a ring around a dissimilar fastener, because the aluminium is being consumed to protect the steel. Both are corrosion, but one is about a coating breach and the other about what was bolted on.
Wipe the area with clean water and a cloth, dry it, and look again later. Then press a fingernail on the edge of the coating around the bloom. If the film is hollow and moves, corrosion has already spread underneath and the visible white is only what escaped. Mark the outline and photograph it so you can compare.
Then look for the origin. A scratch, a stone chip, a cut edge, a thread, a bolt hole, a weld seam. If the bloom radiates from one of those, the breach explains it. If it covers a whole surface evenly with no origin, suspect what the part is cleaned with, because strong alkaline products attack aluminium and strip the anodic layer.
A bloomed part should not be scrubbed with an abrasive pad or a wire brush. That takes off the loose product and the remaining protective layer together, and it spreads the cell along the scratch. A mild detergent, a soft cloth and a thorough rinse are enough to establish whether the surface is stable or still active.
Prevention is a design and process question. Edges and holes need the coating to continue over them rather than stop at them, and fabricated parts need protection restored after welding rather than only applied before it. That is why finish specification matters as much as joint quality in TIG aluminum welding work, and why it applies to every fabricated aluminium part that will see road spray or salt.
