Galvanic corrosion of aluminium requires three conditions at the same time: two different metals, an electrical connection between them, and an electrolyte bridging both. Remove any one and the mechanism stops entirely. That is genuinely the whole of it, and it is why the problem is a detailing problem rather than a material problem.
It is also why galvanic attack is the fastest corrosion mechanism a structure is likely to meet. Pitting takes years to matter. An unfavourable galvanic couple in salt water can consume a fastening in a single season.
Which contacts actually cause trouble
Metals sit in an order of nobility, and the further apart two of them are in that order, the stronger the driving force between them. Aluminium sits towards the active end, which means it is almost always the metal that gets consumed.
| In contact with | Effect on aluminium |
|---|---|
| Copper and brass | Severe |
| Carbon or graphite | Severe |
| Stainless steel | Significant |
| Carbon steel | Moderate |
| Galvanised steel | Protective while zinc lasts |
| Zinc | Protective |
Copper is the one to treat as absolute. Not only does direct contact drive rapid attack, but water running off a copper surface onto aluminium carries dissolved copper ions that plate out and create thousands of tiny couples on a surface that never touched anything. Drainage paths therefore matter as much as physical contact.
Galvanised steel behaves the other way round and is worth understanding properly. Zinc is more active than aluminium, so it sacrifices itself and protects. That protection lasts exactly as long as the zinc does, and when the coating is consumed the underlying steel becomes a cathode and the relationship reverses.
Area ratio decides how fast it happens
The single most useful idea in this subject is that the rate of attack depends on the ratio of cathode area to anode area. A large noble surface driving a small active one concentrates all the current into that small area, and it disappears quickly. A small noble surface against a large active one spreads the same current thinly and the effect may be negligible.
In practice this gives a rule that decides fastener selection. A stainless bolt through an aluminium plate is a small cathode against a large anode and is generally workable. An aluminium bolt through a stainless plate is the same couple with the areas reversed, and it fails. The metals are identical in both cases; only the geometry changed.
Coat the other metal, not the aluminium
The instinct is to paint the metal you are trying to protect. Applied to a galvanic couple this is exactly wrong, and it is worth knowing why. A coating on the aluminium will eventually develop a pinhole, and at that pinhole a very small active area faces the full noble surface, which is the worst possible ratio. The result is a deep, fast perforation instead of a slow general loss.
Coating the noble metal instead reduces the cathode area, which reduces the driving current. A pinhole in that coating exposes a tiny cathode against a large anode, which is the harmless direction. Where only one surface can be treated, treat the more noble one, and where both can be treated, do both.
Non-metals that behave like metals
Three common building materials cause problems that look galvanic and partly are.
Concrete. Fresh concrete is strongly alkaline, and aluminium is attacked by alkalis directly, quite separately from any galvanic effect. Embedded reinforcement adds a second mechanism by providing a cathode. Aluminium cast into or bearing directly on concrete needs a barrier, and the requirement applies to bedding mortar and grout as much as to the structural concrete.
Treated timber. Modern preservative treatments are largely copper-based, and the copper compounds in the timber are available to react with aluminium in the presence of moisture. Decking, bearers and packing pieces in preservative-treated wood therefore need separation from aluminium components, which is a detail that catches out otherwise careful projects.
Carbon fibre composites. Carbon is strongly cathodic to aluminium, and a composite panel bolted directly to an aluminium frame in a wet environment behaves like a graphite electrode. Isolation is mandatory rather than advisable.
What isolation actually looks like
Breaking the electrical path is the most reliable route. Insulating bushes through bolt holes and insulating washers under heads and nuts keep the fastener from touching the aluminium, and a gasket or isolation tape keeps the faying surfaces apart. The detail has to be complete: one metal-to-metal contact anywhere in the joint restores the circuit and the rest of the isolation does nothing.
Breaking the electrolyte path is the second route and it is often easier. Water that drains away leaves no bridge, so drainage, drip details and avoiding horizontal surfaces that hold water do real work. Sealing a joint so water cannot enter is equally valid, provided the seal is continuous, because a partially sealed joint traps moisture instead of excluding it and creates a crevice.
Where neither is possible, sacrificial anodes shift the attack to a component designed to be consumed. They are a maintenance item with a replacement interval, and they belong in the operation and maintenance manual rather than being fitted and forgotten.
What to require in the specification
Name the fastener material and require the area ratio to be favourable. Require an isolation detail at every interface between aluminium and another metal, shown on a drawing rather than described. State that aluminium shall not bear directly on concrete or on preservative-treated timber without a barrier. Require drainage at every horizontal interface. Where anodes are used, require their positions and replacement interval to be documented.
Five requirements, and between them they close off the mechanism that does the most damage in the shortest time. They also apply to fittings added later, which is why they belong in the operation manual as well as in the tender.
Before the details are drawn
Galvanic attack is not a property of aluminium that has to be accepted. It is a consequence of three conditions occurring together, and every one of them is controlled on a drawing. Structures that suffer from it almost never do so because of the design intent; they do so because a fitting was added on site with the fastener that was in the van.
The interface details for every structure we supply are drawn rather than left to the installer, and the certification and design documentation held by DR VERSLAS includes the isolation arrangements so they can be checked and reproduced when anything is added later.