Pretreatment copper parts: same seven tanks, different chemistry

Copper is the material that makes a steel pretreatment line look simple by comparison. A CRCA bracket arrives at the line with mill scale and oil, goes through seven tanks, and leaves with a phosphate film that powder grips. A copper busbar bracket arrives with a 2 nm oxide that has already grown to 5 nm in the bag, a film of drawing lubricant in the stamped pockets, and a requirement that the finished part still conduct where the terminals land. Nothing on the steel sequence answers that, because what pretreatment copper parts need is a different conversion step, a shorter acid window and a cleaner that will not etch the metal.

The failure is not subtle when it happens. Powder goes on, cures, passes a thumbnail test, and lifts from a 40 mm square in sheets when a terminal is torqued onto it, or blisters along the bent edge after one monsoon. The cause is nearly always a step that was skipped because the part looked clean: the pickle, or the DI rinse, or the mask on the contact face.

The sequence itself is degrease, rinse, pickle, rinse, passivate, DI rinse, dry. The numbers below are the ones that hold it together — 5–10% sulphuric acid for 30–60 s, 1–3 g/L benzotriazole at pH 8–9, a final rinse under 10 µS/cm, and a dry oven at 60–80 °C before the part goes anywhere near the booth.

What pretreatment copper parts need that steel parts do not

The first difference is the conversion coating. Iron phosphating works because the acid in the bath attacks the steel surface and releases iron ions, which react with the phosphate to build the crystal layer on the metal itself. Copper releases no iron. The bath attacks the copper, the dissolved copper enters the solution, and what remains on the surface is a loose, patchy deposit that fails a tape test. Zinc phosphating is no better: a zinc phosphate needs a zinc or steel surface to nucleate on, and a copper surface gives it nothing to grow from.

That is not a small inconvenience, because the copper that enters the bath is an oxidiser for everything else in it. In a phosphating solution the accelerator keeps iron in the ferrous state; 10 mg/L of copper ions is enough to push ferrous to ferric, throw sludge, and drop coating weight on steel parts processed later the same shift. A shop that runs copper through the same tank as its steel work finds its phosphate coating weight falling and its sludge rising, and blames the chemical supplier.

The second difference is the passivation chemistry. Steel is passivated with a chromate-free rinse or left to the phosphate; copper tarnishes in hours in Mumbai humidity because the oxide keeps growing. The answer is an organic passivate, benzotriazole being the common one, which chemisorbs onto copper as a Cu(I)-BTA complex and blocks further oxidation. It is 5–50 nm thick, it is hydrophobic, and it is also a dielectric — which matters as soon as the part has an earth stud.

The third difference is drying. Copper shows water spots more than steel because the dissolved salts in a rinse droplet concentrate as the droplet evaporates and leave a visible ring on a warm-toned surface. A final rinse under 10 µS/cm and a 60–80 °C air dry remove the problem; a drag-out rinse and an air knife do not.

The seven stages, mapped onto a copper part

The table below is the sequence we specify for copper and brass before powder, with the operating window and where each stage differs from the steel route.

Degreasing copper without etching it

The cleaner matters more on copper than on steel because the safe window is narrower. A silicate-free alkaline cleaner at pH 9.5–11 removes stamping lubricant and fingerprint soil in 3–5 minutes at 55–60 °C without attacking the metal. Above pH 11 with a strong chelant the surface darkens, and the darkening does not come off in the pickle, so it shows through a thin powder film as a brown patch.

Two ingredients to keep out. Ammonia and ammoniacal cleaners dissolve copper through the cuproammonium complex, which is why a cleaner borrowed from a brass polishing line can remove 0.5 µm of surface in a single dip and leave a matte, grainy finish. Chloride is the other: chloride is the classic pitting initiator for copper, and a cleaner or rinse carrying 30 mg/L or more leaves the part vulnerable under any film that is later scratched.

Stamped copper parts also carry copper fines. A blanking operation on a 250 T press leaves swarf pressed into the surface at the die entry, and that swarf is the same colour as the part until the coating goes on and it oxidises green under the film. A 30-second ultrasonic clean or a brush in the degrease stage is the difference between a clean part and one with a green freckle that appears three months later.

Pickling: the step that decides adhesion

The pickle does two things: it strips the oxide and it activates the surface for the passivate. Five to ten percent v/v sulphuric acid at ambient for 30–60 seconds is enough for the oxide, which is nanometres thick and dissolves in seconds. The time limit exists for the metal underneath, because the acid keeps dissolving copper once the oxide is gone.

A worked number shows why the limit matters. Copper dissolves in 5% sulphuric acid at roughly 1 µm/min at ambient with no inhibitor. Sixty seconds removes about 1 µm of metal, which is 8.94 g/m² at a density of 8.94 g/cm³. For a bracket with 0.12 m² of wetted area, that is 1.07 g of copper per part, and 600 parts put 640 g into a 1,000 L tank — 640 mg/L, sixty times the 10 mg/L that would ruin a phosphate bath, and enough to make the pickle itself slow down as copper builds up in it. The bath is dumped on copper loading, not on a calendar.

Brass needs a tighter limit than copper. Above about 60 seconds in 5% acid, a 63/37 brass begins to lose zinc selectively from the surface, leaving a pink, porous copper layer that looks like a plating defect and bonds poorly. Keep the dip short, keep the acid cool, and if the part is a machined brass fitting with a leaded grade, expect a smut that needs a 10-second dip in a dilute inhibited descaler rather than a longer acid soak.

Passivation, and why it is not paint

Benzotriazole at 1–3 g/L, pH 8–9 and 40–50 °C for 30–60 seconds builds the Cu(I)-BTA film by chemisorption: the molecule attaches to the copper through its nitrogen atoms and the ring stands up from the surface. The film is 5–50 nm, depending on concentration and time, and it slows oxide growth enough to keep a copper part bright for months indoors.

Three practical points. First, pH control is the whole process: below pH 7 the BTA does not adsorb properly, and above pH 10 the copper starts to etch. A dosing pump and a daily pH check, not a weekly one, because the drag-out from the pickle moves the tank in a shift. Second, the film is a dielectric, so it must not reach a contact face — mask plugs and tape go on before the part enters the line, not after coating. Third, the passivate is not a substitute for the pickle: BTA on an oxidised surface seals the oxide in, and the powder then adheres to the oxide through the film.

If the drawing does not call for a passivate, the copper will still take powder after a pickle and a DI rinse, provided the part goes to the booth within a shift. That is the practical alternative where a customer wants a plain industrial finish and no additional chemistry, and it is worth stating on the drawing so the shop does not add a step nobody is paying for.

Drying, curing and the 200 °C problem

Drying at 60–80 °C with forced air takes 10–15 minutes on a jigged part. The jig matters: a copper bracket with a 12 mm blind hole holds water that flashes in the oven and leaves a stain ring around the hole. Jig it hole-down, or blow it off with compressed air before the oven.

The oven that cures the powder runs at 180–200 °C for 10–15 minutes, and copper oxidises at that temperature even under a film. The oxide thickens from nanometres to tens of nanometres and the bare copper at a masked face goes brown. On an industrial bracket nobody notices. On a decorative brass panel with a 3 mm masked border, the border comes out a different colour from the panel the customer approved, and that is a colour-match argument rather than a coating failure.

Thermal expansion only becomes a design issue in mixed-material assemblies. Copper expands at about 16.5 × 10⁻⁶/K against 11.7 × 10⁻⁶/K for steel. Over a 155 K cure on a 400 mm copper busbar bolted to a steel frame, the differential is (16.5 − 11.7) × 10⁻⁶ × 400 × 155 = 0.3 mm. A powder film can follow that; a rigid two-pack epoxy at 150 µm on a joint that is then restrained cannot, and the film cracks at the fastener.

Proving adhesion on a non-ferrous substrate

Three tests release a coated copper part. Dry film thickness comes first, and it has to be measured with the right instrument: a magnetic induction gauge needs a ferrous base, and on copper it either errors or reads zero. An eddy-current gauge calibrated on an uncoated copper offcut from the same batch is the correct tool, and the calibration matters because copper conductivity varies with alloy and temper.

Adhesion is checked by cross-cut to ISO 2409, which specifies the lattice spacing by film thickness: 1 mm for films up to 60 µm, 2 mm for 60–120 µm, and 3 mm above that. A 90 µm film cut at 1 mm spacing is not an ISO 2409 test even if the knife is sharp. Gt0 or Gt1 is the acceptance band; Gt2 with flakes at the intersections is a reject on a part that will be handled.

Salt spray is the third, and it is where copper parts are most often misgraded. ASTM B117 runs the neutral salt fog; ISO 4628-3 grades the result by degree of rusting, and rust is an iron phenomenon. On a copper substrate there is no red rust to grade, so the assessment has to move to blistering at the scribe, creep from the scribe, and a colour change on the coated face. A specification that says "480 h, no red rust" on a copper part passes any film that stays on, including one that has blistered over half the panel.

The line, and what it will not do

Copper and brass parts run through the same plant as our steel work at Ganesh: cut on the 12 kW fibre laser or stamped on presses from 20 T to 300 T, formed on the 200 T, 5-axis brake, degreased through the seven-tank pretreatment, and coated in the 18 ft × 14 ft electrostatic powder booth with an in-house curing oven. DFT and cross-hatch adhesion are checked in the paint shop on every batch, and the records travel with the parts under our ISO 9001:2015 system. The copper and brass work sits alongside the aluminium fabrication in copper, brass and aluminium fabrication, and the coating route in the paint shop.

What we do not do is pass copper through the iron phosphate tank. That stage belongs to the steel and aluminium parts on the line, and a copper part in it damages the bath and the part at the same time. Salt spray testing is not run in our paint shop either: it is an external laboratory test, quoted per submission, and the first article is the right place to spend the money. Chromate conversion is not offered because it conflicts with RoHS.

The tank-by-tank sequence for the steel parts on the same line is in seven-tank pretreatment, process explained, the degreasing chemistry that precedes it in degreasing before powder coating, and the reason a phosphate film exists at all in phosphating and conversion coatings for steel. All coating guides are indexed under surface treatment and powder coating.

What to put on the drawing

1. Alloy and temper, not just "copper": C11000, C26000 brass and a tellurium copper are three different pickling problems.
2. Which faces conduct, and therefore which faces are masked before pretreatment.
3. The passivation requirement, or a statement that none is required.
4. Powder type, colour, gloss and DFT, with the indoor or outdoor band.
5. Adhesion test method and acceptance class, and whether the part is cross-cut on the first article only.
6. Salt spray hours, the assessment criteria, and whether the scribe is included.
7. Handling and packing, because a passivated copper face marks from a bare hand within a day.

When coating copper is the wrong answer

A busbar joint does not get coated. Neither does an earth stud, a sliding contact, or a soldered terminal land. Where the copper is there to conduct, the answer is masking or a selective plating such as tin, which is bought in and does not involve the powder line at all.

A part that will sit in a sulphur-rich atmosphere, against a rubber gasket that releases sulphur, or in a coastal switchroom, is better served by a heavier film than by a thin decorative coat: 100–120 µm over a pickled and passivated surface, with the fasteners specified in stainless rather than zinc-plated steel, because a copper cathode next to a zinc anode in a damp enclosure sacrifices the fastener first.

Send the alloy, the masked faces and the finish with the drawing. Request a coating quote, and the powder coating cost estimator gives a first figure for film area and powder weight while the drawing is still moving.