A blast profile can look perfect and still fail. Why chloride and sulfate salts must be measured before the first coat, and how the Bresle test works.
A steel surface can pass every visual check — Sa 2½, sharp angular profile, no rust, no mill scale — and still blister within a few years. The reason is usually invisible: soluble salts left in the bottom of the profile. Blasting removes oxide and scale, but it also drives contamination deeper into the pits it creates, and if those salts are not washed out, the first coat goes on over a layer that will pull water to the interface for the rest of its service life.
Where the salts come from
- Marine atmosphere. Chloride settles on steel near the coast and inside shipyards. A structure 200 m inland is not the same substrate as one on the quay.
- De-icing salt and process spillage. Bridge decks, plant floors and tank farms carry salt spray and chemical drips that look like ordinary dirt.
- Residues under old coatings. A coating that has been in service for fifteen years has concentrated salts at the interface. Removing the paint exposes them.
- The abrasive itself. Recycled grit and sea-dredged sand can carry chloride into the surface faster than blasting takes it out. Abrasive cleanliness is part of the salt problem, not a separate subject.
Why a thicker film does not fix it
Salt at the interface creates a concentrated electrolyte in a very small volume. Water vapour from the air permeates the coating, dissolves the salt, and the resulting osmotic pressure pulls in more water — the classic blister. Chloride also breaks down the passive layer on steel locally, so pitting can begin under a film that still looks intact. Neither mechanism depends on the coating being thin: a 300 µm high-build epoxy blisters over a contaminated interface much as a 100 µm primer does. The defect sits at the boundary, and the boundary is the one place a coating cannot protect.
Measuring it
The field method is the Bresle patch. A fixed area of the surface is sealed, a known volume of distilled water is injected and extracted, and the conductivity of the extract is converted into surface salt density. ISO 8502-6 defines the extraction and ISO 8502-9 the conductometric determination; results are normally reported in mg/m² as sodium chloride equivalent.
Where the tests are taken matters as much as the method. Salts concentrate where water sits and where the steel was hardest to clean: weld toes, the inside of stiffeners, the bottom third of a tank, areas the blaster could not reach properly. An average over a whole structure hides exactly the locations that will fail first. Acceptance limits in project specifications commonly fall in the 20–50 mg/m² range for sodium chloride equivalent, tighter for immersed service and for ballast tanks. The figure that governs is the one in the specification and the coating manufacturer's written guidance — publicly available manufacturer materials and the project specification should agree before work starts.
When the reading is too high
Washing is the only real remedy, and it has to be done properly: high-pressure fresh water, with a wetting agent where the specification permits one, worked into the profile rather than swept across it, followed by drying and a repeat measurement. Two details decide whether the wash works. The water must be clean — potable or deionised, not harbour water. And the surface must be dry before coating, because a damp profile re-concentrates salts at the bottom of the pits as it dries. If the profile has already flashed rust, the sequence is wash, dry, light sweep blast — not a heavier coat.
Writing it into the specification
Put the salt limit, the test method, the number and location of tests and the action on failure into the specification, not only into the method statement. State that testing happens after blasting and before the first coat, and again after any wash. Zinc-rich primers such as Interzinc 52HS and high-build epoxies such as Intergard 475HS both go directly onto the blast, so both sit on whatever the blaster left behind. A salt clause protects the whole system, not just the first coat.
A Bresle test takes a few minutes. A fresh-water wash on a bridge or a tank costs a day of production. Blistering found at a five-year inspection costs full removal and re-coating, in a place that is harder to reach than it was the first time. Industry failure analyses keep putting surface preparation at the top of the list of causes, and soluble salts are one of the recurring items on it. That is a cheap test to specify.