Corrosion is tricky and nearly impossible to rationalize, less so with simple comparisons like redox potentials. Experimental
data is the way to go. On the other hand, wrong ideas can survive decades in the most reputable sources: at buildings, think of radioactive tips in lightning rods, recommended and used for a century, dismissed recently.
I'm not really convinced by the effect of galvanic couples in alloy corrosion. Few quick experiments I did showed no action. Possibly it makes all the difference in batteries, where the metals are very pure, and little in constructions, where we use alloys that contain already precipitates that build corrosion couples within a single part.
One very important distinction among the elements and alloys is whether they build a protective oxide layer
. If not, simple theories like the redox potential may apply more or less. But if the oxide layer protects the metal, the integrity of this layer decides everything and the redox potential very little.
So you may sort mentally the metals and alloys in three groups:
- Noble metals resist corrosion because their oxides (and hydroxides) aren't very stable. Especially, their oxide is less favourable than the oxide of hydrogen, water. Upon contact with rainwater or air moisture, water's oxygen atom remains at the hydrogen and doesn't switch to the metal.
- Metals that resist corrosion by a protective oxide layer.
- The other metals don't resist corrosion.
Most metals don't resist corrosion. Sodium, calcium... will corrode. Iron is at the limit of this category.
The noble metals include gold, palladium, platinum... Expensive for buildings, but perfect for electric contacts as a thin layer. Nickel, tin are about the only affordable metals here. Copper and lead are at the limit of this category. In this category, the redox potential is an indication of the corrosion behaviour.
The metals with a protective layer include tantalum, chromium, aluminium, silicon (...for chemists not a metal), titanium, zinc. Their redox potential can look very unfavourable but they may resist corrosion perfectly, the best examples being tantalum and titanium. Most corrosion-resistant alloys affordable in buildings are in this category. Copper and lead are at the limit of this category too.
Alas, the build-up and integrity of the oxide layer is about impossible to predict. That's very much observation and afterwards explanation.
As an example, over 13% chromium suffice in iron to build the oxide layer and make steel stainless (so this is not a matter of redox potential). Though, chloride ions harm the layer, so seawater operation demands special stainless steel. And if the steel with chromium contains too much carbon, they precipitate as chromium carbide, especially during welding, leaving too little chromium in the matrix to be protected, so very low carbon or more chromium is the answer.
As an other example, pure aluminium resists corrosion thanks to its oxide layer, but Al-Cu and Al-Zn alloys don't. Al-Si and Al-Mg are good. These would be the ones where the alloying elements don't precipitate and where heat treatment has no hardening effect, but Al-MgSi makes precipitates, hardens by heat treatment, and resists corrosion nevertheless. Tiny additions, under 1%, can spoil the corrosion resistance, which also depends on the corroding medium, like sulphur oxides in the rain.
And so on and so forth. I see no hope to predict the effect of galvanic couples here.
Oxide layer isn't the only option. An iron phosphate layer too can protect steel somewhat.
Anode and cathode: chemists seem to stick with the original definition, all other people understand them differently, including the manufacturers of accumulators, so be prudent. As for the current flow, it makes a loop in the source and load, so if flowing from + to - in the load, it's the opposite in the source, and a battery or a corrosion couple is a source.