What a Coating Is Protecting Against

Why this matters

Coating gets treated as a single idea: something got painted, so it is protected. It is at least three different ideas that fail in three different ways. The most expensive version of the mistake is not leaving something bare. It is coating the wrong member of a dissimilar-metal pair, which concentrates all the corrosion that used to spread over a large area into a few pinholes and drives them through the wall in a fraction of the time bare metal would have taken.

Before you disturb a coating to look under it

Scraping, sanding, grinding, or heating a coating is where the coating stops being a corrosion topic and becomes an exposure topic.

• Presume lead on pre-1978 buildings. For compensated renovation, repair or painting that disturbs painted surfaces in target housing or a child-occupied facility built before 1978, above the rule's minor-repair thresholds, the EPA Renovation, Repair and Painting rule at 40 CFR 745 Subpart E requires a certified firm, a certified renovator, and contained work practices. Test the paint or presume lead. Do not dry-sand or open-flame it either way.
• Treat an unknown yellow-iridescent or dull plating on legacy hardware as cadmium. Heating, welding or cutting cadmium plating produces cadmium oxide fume, which is acutely toxic and has no useful warning properties: it does not smell dangerous and the symptoms arrive hours after the exposure that caused them. Do not apply heat to it without local exhaust ventilation or supplied-air respiratory protection under 29 CFR 1910.1027. If you cannot identify the plating and cannot ventilate, remove the hardware cold and replace it.
• Heating or cutting zinc-coated steel releases zinc oxide fume. Galvanizing is zinc, and zinc oxide fume causes metal fume fever. Welding, cutting or brazing on zinc-bearing base metal needs local exhaust ventilation or airline respiratory protection, which is the requirement in 29 CFR 1910.252(c) for welding, cutting and heating. Grinding the zinc off the weld zone first and ventilating is the ordinary field answer.
• Treat an unknown industrial primer as chromate-bearing. Grinding or blasting one can generate hexavalent chromium dust, and hexavalent chromium is an INHALATION carcinogen, so the controlling protection is respiratory and not gloves. Use wet methods or local exhaust extraction rather than a wire wheel in still air, with respiratory protection selected under 29 CFR 1910.1026, plus eye and face protection, and the glove class named on the stripper's own safety data sheet, which must be available under 29 CFR 1910.1200.

Three mechanisms

Barrier. Paint, powder coat, most platings, anodizing, tank and pipe linings. The film keeps electrolyte off the metal. It has no chemistry of its own to offer and it is only as good as its continuity. Every break in the film is a break in the protection at that spot.

Sacrificial, also called galvanic. Hot-dip galvanizing, zinc-rich primer, zinc and aluminum thermal spray, bolt-on anodes. A metal anodic to the substrate is put in electrical contact with it and corrodes preferentially. The film does not have to be continuous, because the surrounding coating protects a small break in itself.

Inhibitive. Conversion coatings and primers carrying inhibitive pigments. They chemically passivate the surface, or slowly release a species that does, so a small break heals rather than propagates. They are usually a layer in a system rather than the whole system.

The single most useful contrast in the whole subject: a scratch through paint on steel is a wound, and a scratch through galvanizing is covered. The paint scratch rusts, and the rust creeps under the film from the edges, which is why a chip on a painted panel grows into a patch. The galvanizing scratch sits in a field of zinc that goes on protecting the exposed steel at the bottom of it for as long as the surrounding zinc lasts.

The area rule

Corrosion current in a galvanic couple is driven by the potential difference between two metals, but where that current lands is set by area. The anode, the less noble metal, dissolves. If the anode is small and the cathode is large, the current density concentrates on a small patch and penetrates fast.

That is why a broken barrier coating over a noble surface is worse than no coating. Nickel or chrome plating on steel is more noble than the steel under it, so a pinhole through the plating is a tiny anode of exposed steel facing a large cathode of plating. The plating did not cause the corrosion, it concentrated it.

The rule that falls out of this is counterintuitive and worth memorizing: if you can only coat one member of a dissimilar-metal joint, coat the cathode, the more noble metal. Coating the anode leaves small holidays that become tiny anodes against a large cathode. Coating the cathode shrinks the effective cathode, and a holiday in it is just a small cathode, which is harmless.

Coating families at a glance

What a coating is not protecting against

This is where most of the field mistakes live.

• It does not break a galvanic couple. If the two metals remain in metallic contact under the coating, the couple exists and the current still flows. Breaking the couple means an isolating sleeve, washer, or dielectric fitting, not paint.
• It does not restore lost wall. A coating over a pitted pressure boundary buys appearance, not integrity. Pitting is a replace, not a repaint.
• It does not stop water that is already getting in. Coating over a wet or contaminated substrate traps the moisture, and the film blisters from beneath. The leak gets fixed first or the coating is wasted.
• It does not protect a cut end, a drilled hole, or a thread you made on site. Every field cut through a coated product exposes bare substrate. Galvanized steel cut in the field gets a cold zinc-rich repair on the cut end, which is a specified step in trade practice rather than an optional nicety.
• It does not survive being a ground or bonding path. Anodizing and paint are insulators, so a bonding connection made onto a coated surface may read fine on a meter and fail under fault current. Bonding surfaces get masked before coating, or scraped clean and re-protected around the connection after.
• Paint over rust does not stop rust. The moisture and the corrosion products are already there and the reaction continues under the film.
• It does not raise a temperature limit, and it does not make a material chemically compatible with what it carries. Only a lining qualified for that fluid and that temperature does the second one, and one holiday in a lining is a full-strength attack at one point.

One couple, five options

A stainless steel bracket bolted through a cast aluminum housing, outdoors, in coastal air. Stainless is strongly cathodic to aluminum, so the aluminum is the anode and the aluminum is what will be eaten.

Take the wetted contact areas as illustrative round numbers, because the reasoning is about the ratio rather than the values: roughly 20 square inches of stainless in the wetted zone against roughly 0.5 square inches of aluminum exposed at the joint. That is a 40 to 1 cathode-to-anode ratio before anyone paints anything, which is already unfavorable.

Option 1, coat the aluminum only. This is what most people reach for, because the aluminum is the part that corrodes. Say the coating goes on well and 2 percent of the aluminum's coated area ends up as holidays at edges and fastener contact points. The exposed aluminum drops from 0.5 square inches to 0.01 square inches while the stainless cathode stays at 20 square inches. The same driving current now lands on one fiftieth of the area it used to, because 0.5 divided by 0.01 is 50, so the local penetration rate at those holidays goes up roughly fifty-fold. This option is worse than doing nothing, and it looks like the responsible choice.

Option 2, coat the stainless only. With the same 2 percent holiday assumption, the effective cathode drops from 20 square inches to 0.4 square inches against the unchanged 0.5 square inches of aluminum. The ratio goes from 40 to 1 down to 0.8 to 1, and the aluminum's corrosion rate falls with it. This is the counterintuitive answer, and it follows directly from the area rule.

Option 3, coat both. Better than either single option, and what a coating specification would call for. It still leaves a couple that is only managed, so its life is the life of the weaker film.

Option 4, isolate. A nonconductive sleeve through the bolt hole, isolating washers under the head and nut, and a sealant that keeps the crevice dry. There is now no metallic path between the two metals, so there is no couple to manage. This removes the mechanism rather than slowing it, and it is the answer wherever the joint can accept the extra stack height.

Option 5, change one metal. A fastener and bracket closer to aluminum in the galvanic series. Best where you are specifying, useless where you are standing in front of an existing assembly.

Field order of preference: 4, then 3, then 5 where you get to specify, then 2 if that is all you can do. Option 1 is the one to actively unlearn.

Prep decides the outcome more than product does

Coating failures trace to surface preparation far more often than to the wrong product. Three prep conditions do most of the damage:

• Soluble salts left on the surface. Chlorides on a coastal or road-salt asset pull moisture through the film by osmosis and lift it from beneath. Washing before abrasive prep, not after, is what removes them.
• Silicone contamination. Silicone lubricants and release agents cause craters and fisheyes and are stubborn to remove, which is one more reason to know what lubricant has been used on an asset before it gets coated.
• Coating onto a surface at or near dew point. Standard trade practice is to apply only when the substrate is at least 5 degrees Fahrenheit above the dew point and rising, because an invisible film of condensation on cold steel puts water under the coating at the moment of application.

Reading which mechanism you are standing in front of

You will often arrive at an asset with no record of what is on it. Two checks separate the families, and neither needs a lab.

• Look at an existing scratch or chip, not a fresh one. Rust bleeding out of the break and creeping under the film at its edges means barrier over steel with no sacrificial layer. A break gone dull grey and stayed dry, with no rust staining below it, means zinc is still working nearby.
• Look at the cut ends and drilled holes. A specified galvanized assembly usually shows a different texture where field cuts were touched up. Untouched bright cuts on an otherwise dull grey product mean nobody repaired them, which is where that asset will fail first.

Judging whether a coating has actually failed, as opposed to identifying what it is, is a separate read with its own signatures and rating scales. See the sibling article on telling whether a coating has failed.

References

• 40 CFR 745 Subpart E, the EPA Renovation, Repair and Painting rule, for lead-safe work practices when disturbing paint in pre-1978 target housing and child-occupied facilities
• 29 CFR 1910.252(c), ventilation and protection in welding, cutting and heating, for zinc-bearing base metal
• 29 CFR 1910.1200, hazard communication, for safety data sheets on coatings, strippers and solvents
• Manufacturer product data sheets for substrate temperature, dew point margin, and required surface profile
• See related: How to Tell Whether a Coating Has Failed; The Galvanic Series as a Field Tool