Galvanic corrosion in mixed-material assemblies

Two metals in one assembly form a cell. The electrolyte is the condensation that forms on a canopy at 5 a.m., the circuit closes through a bolt and a bracket, and the paint break at the hole is where the current leaves. Every material choice is correct on its own. The assembly still pits through at the fastener, because galvanic corrosion needs only three things: two metals, an electrolyte and a path between them.

What sets the rate is not the metal pair alone. It is the wetted cathode area divided by the wetted anode area, multiplied by the current the couple drives through the electrolyte. A 4 cm² stainless fastener in contact with 1 cm² of wetted aluminium in a chloride atmosphere can dissolve that aluminium at about 0.87 mm per year. A 3 mm bracket has nothing to spare.

The sections below give the galvanic series as it is used on a shop floor, the Faraday arithmetic that turns a potential difference into millimetres lost per year, the joint details that break the couple, and the point where a mixed-material joint should be redesigned rather than protected.

Which metal becomes the anode in galvanic corrosion?

The less noble metal in the galvanic series is the anode and loses section. The more noble metal is the cathode and is protected. Zinc is the anode in a zinc-plated bolt against a CRCA bracket, aluminium is the anode against stainless, and mild steel is the anode against 304.

The potentials below are approximate, measured against a saturated calomel electrode in seawater. ASTM G82 sets out how a galvanic series is built and how it should be read, and the reading matters more than the digits: the series ranks metals, it does not give a rate. Polarisation, area ratio, conductivity and time of wetness decide whether a 0.7 V couple removes 2 µm a year or 0.9 mm.

The last row is the one that catches lift frames and genset skids. A galvanised frame with stainless fasteners is a 1 V cell, and the zinc at the fastener is small compared with the stainless it faces. The frame is not under-protected. It is being consumed at the fixing, and the corrosion product arrives as red rust on a surface that was specified to need no maintenance for a decade.

The series also has a trap in the middle. Zinc is anodic to aluminium by about 0.24 V, so a zinc-plated bolt in an aluminium panel protects the aluminium while the plating lasts and then reverses into an aluminium-versus-steel couple once the plating is gone. A joint that looks fine for two years and fails in the third is often this sequence rather than a single couple.

How does the area ratio set the damage rate?

Corrosion removes metal in proportion to current, and current is set by the cathode area. Faraday's law gives the conversion directly: 1 A/m² dissolves 1.16 mm of steel, 1.09 mm of aluminium or 1.50 mm of zinc per year.

The current density on the anode is the cathode density multiplied by the area ratio:

i(anode) = i(cathode) × (A(cathode) / A(anode))

That one line explains most mixed-material failures in a fabricated assembly. A stainless M6 bolt with 4 cm² exposed, coupled to 1 cm² of wetted aluminium inside a clearance hole, runs a 4:1 ratio. If the couple drives 20 µA/cm² on the stainless, the aluminium in that crevice sees 80 µA/cm², or 0.8 A/m², which is 0.87 mm/yr. The crevice is also where the electrolyte sits longest and where oxygen is least available, so the loss concentrates in the one place that cannot be inspected.

Reverse the couple and the failure gets faster. An aluminium rivet in a stainless panel presents an anode of about 0.4 cm² to a cathode measured in square metres. The ratio runs into the thousands, the rivet head disappears, and the panel is left with an elongated hole and a stain. That is why aluminium rivets belong in aluminium panels and nowhere else.

A galvanised bracket with stainless fasteners runs the same arithmetic in the other direction. The zinc around each fastener is a small anode, the stainless is a large cathode, and the current density on the zinc is high. At 10 µA/cm² the zinc dissolves at 0.15 mm/yr, so a 20 µm coating is gone after seven weeks of continuous wetting. Wetting is rarely continuous, which is why the part lasts a year rather than two months, and why the failure still arrives before the maintenance interval.

Where does the electrolyte come from in a dry plant?

A couple needs a film of water to run. In a sealed cabinet at 40% relative humidity the film is a few molecular layers and the current is negligible. The same joint in a coastal shed crosses the threshold every night.

ISO 9223 classifies atmospheric corrosivity into categories C1 to CX using time of wetness, sulphur dioxide deposition and airborne salinity. The category is a better guide than the word outdoor, because C3 industrial and C4 coastal differ by a factor of two in steel loss and by much more in chloride load.

The practical sources of that film, in the order they cause failures:

• Condensation on a part that cools below the dew point overnight, which is most outdoor fabrications in Mumbai between June and September.
• Trapped water in an upward-facing channel, an open hem or a lap joint with no drain.
• Wash-down, which puts chlorinated water into every crevice once a week.
• Airborne chloride at a coastal site, which deposits as salt and then holds moisture at 75% RH rather than 100% RH.

The design controls are cheap. A 6 mm drain hole at the low point of every closed section, a 3° fall on a horizontal face, no upward-facing ledge narrower than 20 mm, and a sealed joint face rather than an open lap. None of them adds a process step, and each removes a reservoir.

Which fasteners and joints break the couple?

Matching the fastener to the panel is the first control and the cheapest. Stainless bolts in a stainless frame, zinc-plated bolts in a CRCA bracket, aluminium rivets in an aluminium panel. A matched joint has no couple and needs no isolation, which is why it is also the most reliable.

Where the materials cannot match, the joint has to be electrically broken. A nylon or PTFE washer under the head, a sleeve through the hole and a sealing washer on the far side isolate the fastener completely, leaving two insulated faces and a mechanical clamp. The check is a resistance measurement across the assembled joint with a 500 V insulation tester, which should read above 1 MΩ. A joint that reads a few ohms has a metal-to-metal bridge at one point, and that point will corrode.

Three details decide whether the isolation survives the first year:

The washer has to outlast the paint. A 0.5 mm nylon washer creeps under bolted preload and relaxes within months at 60 °C. A fibre-reinforced or PEEK washer holds its thickness. The preload belongs on the drawing, because an over-torqued isolated joint crushes the washer and shorts the couple.

The sealant has to be neutral-cure. Acetoxy-cure silicone releases acetic acid as it cures, and acetic acid attacks zinc and aluminium. A joint sealed with the wrong silicone corrodes faster than an unsealed one, because the acid is trapped against the metal.

Anti-seize has to be graphite-free. Graphite is cathodic to steel and to zinc, so graphite-loaded anti-seize on a stainless bolt turns the paste into a cathode. Molybdenum disulphide or a wax-based compound avoids that, and stainless-on-stainless threads need one of them to prevent galling.

Where the joint is structural and the environment is wet, a zinc-rich primer at 40–60 µm on the steel face gives sacrificial protection at the contact. It is a coating, not a battery, and it works only if the contact face is fully covered before assembly.

Welded joints deserve a separate mention, because they are usually assumed to be immune. A weld between two dissimilar metals is itself a galvanic couple, and the weld metal, the heat-affected zone and the parent metal sit at three different potentials. We do not weld aluminium to steel. Where a drawing shows that joint, it is a mechanical or adhesive joint, and it should be drawn as one.

What does Ganesh run, and where is the couple decided?

Mixed-material joints reach us as drawings, and the fastener callout is the line we cannot correct after the fact. The material side is in-house: 5052 and 6061 aluminium cut on the 12 kW fibre laser and formed on the 200 T, 5-axis brake with a 4,450 mm bed, and steel parts through seven-tank pretreatment and an 18 ft × 14 ft powder booth with a curing oven.

That matters because the coating is the main lever on the anode. A part cut, deburred, phosphated and coated on all faces has a different corrosion life at the joint than one coated on the visible side only. Where the drawing calls for a coated contact face, the part is coated before assembly and the fastener is isolated at build.

What we do not do is weld aluminium to steel, hot-dip galvanise in-house or apply zinc flake coatings. A joint that needs any of those is a different route, and the quotation should say so rather than absorb it.

The materials and grades hub carries the substrate side of this decision. The 304 stainless corrosion guide covers what the stainless fastener is doing electrochemically, 5052 aluminium corrosion and finish compatibility covers the anode, and the coastal site corrosion guide sets the exposure side. The paint shop capability page shows what the coating line covers, and custom parts lists the material range we cut and form.

What to put on the drawing

• The couple named as a joint, not left to the fastener schedule: "M6 × 20 A4-70 bolt, nylon washer and sleeve, isolated from panel".
• The isolating component and its material, with the preload torque that does not crush it.
• Which faces are coated, and which faces must be bare for earthing.
• The drain hole diameter and position on every closed section.
• The sealant chemistry, neutral-cure, and the joint face it goes on.
• The exposure category to ISO 9223, so coating and isolation are sized to the same environment.
• For earthing points, the star washer and the bare face, kept away from any aluminium or zinc surface that will be wetted.
• A salt spray requirement to ISO 9227 on an assembled or simulated joint, with the hours and the acceptance criterion.

When a mixed-material joint is the wrong answer

Some joints should be redesigned, not protected. A 1 V couple in a C4 or C5 atmosphere with continuous wetting will consume a zinc coating faster than any maintenance interval, so the answer is a different substrate, not a better washer. The same applies where the anode is small and hidden: an aluminium rivet in a stainless panel cannot be inspected and cannot be replaced without destroying the joint.

If the assembly is immersed, or the design life is longer than the coating system, the material has to change. 316L instead of 304, aluminium-only construction with aluminium fasteners, or a plastic isolator that is designed in rather than added at build.

The other case is the one that gets argued about: a bare stainless panel bolted to a bare galvanised frame on a C3 indoor site, where both surfaces are visible and neither can be coated. There is no washer that fixes that. Change the frame to painted CRCA, change the panel to 304 with zinc-plated fasteners, or accept a repaint interval.

Send the drawing with the joint details and the exposure category to request a quote, and the DFM review will flag the couples the drawing has left open.