Key takeaways
- Copper carries a 1–3 nm native oxide that regrows within an hour of cleaning and thickens in the cure oven, so the window between the final rinse and the powder booth is measured in hours, not days.
- Use a non-etching alkaline cleaner at pH 9–11 and 45–55 °C for 3–5 minutes; acidic and ammoniacal cleaners attack copper, and ammonia can stress-crack stressed brass.
- Copper does not belong in a zinc phosphate tank: copper ions poison the bath and the part gets no useful conversion film, so run copper on a separate cycle and never share the tank.
- A 500 mm copper panel grows 0.4 mm more than its steel frame over a 180 K cure cycle, so specify a flexible coating and avoid coating a rigidly constrained joint.
- Read salt spray on copper as scribe creep and blistering, not red rust; copper corrosion products are green or white and a red-rust acceptance criterion will pass a failing part.
One gram per square metre of drawing oil is a 1.1 µm film at 0.9 g/cm³, and on copper it is invisible against the metal's own colour. Degreasing copper parts is not a harder job than degreasing steel; it is a different one, because the substrate underneath the oil is already reacting with the air. Copper carries a native oxide a few nanometres thick, that oxide regrows within an hour of cleaning, and it grows again during the cure cycle. Powder bonds to whichever surface it finds, and copper oxide is a weak boundary layer.
The second difference is chemical. Copper is attacked by acids and by ammonia, so the strong alkaline and acidic stages that suit CRCA are the wrong tools. The third is physical: copper expands about 16.5 µm per metre per kelvin against 12 for steel, so a coated copper panel on a steel frame moves under the film every time the assembly heats and cools.
This page sets out the cleaning cycle that works on copper and brass, the coating window, the thermal numbers that decide the coating choice, and the tests that actually mean something on a copper substrate.
What is different about degreasing copper parts?
Three things: the oxide, the chemistry the metal tolerates, and the absence of a useful conversion coating. Steel pretreatment is built around removing oil and then growing a phosphate crystal on the surface. Copper offers neither half of that model.
The oxide is the first problem. A clean copper surface oxidises in air at room temperature, reaching 1–3 nm within minutes and continuing to grow with humidity and time. Heat accelerates it sharply: at 180–200 °C, the metal temperature of a powder cure, a bare copper surface tarnishes visibly in seconds. If the part is coated with a 10 nm oxide at the interface, the powder adheres to the oxide rather than to the metal, and the failure appears as edge lift or a blister that grows from a scribe.
The chemistry is the second. Copper dissolves in mineral acids and in ammonia, and copper alloys are sensitive to both. A cleaner at pH 2 that strips oxide from steel will etch copper and leave a smut; an ammoniacal cleaner that removes tarnish from brass will, on a stressed part such as a formed busbar, raise the risk of stress-corrosion cracking. Strong chelating agents hold copper in solution and can leave the surface active and unstable.
The third is the conversion coating. Copper does not form a zinc or iron phosphate film of any use, and the copper that dissolves into a phosphating bath contaminates it: copper ions in a zinc phosphate tank produce a poorly adhering, dark coating on the steel parts that follow. A line that shares tanks between copper and steel is a line that will produce a mystery adhesion failure on steel within a week.
The cleaning line for copper, stage by stage
The sequence below is the working route for copper and brass parts from 0.5 mm sheet up to 10 mm plate, coated with polyester or epoxy powder.
| Stage | Typical condition | What goes wrong | Ganesh limit or note |
|---|---|---|---|
| Incoming check | Oiled, tarnished or laser cut | Mixed surface conditions in one rack | Sort by surface condition before loading |
| Degrease | Non-etching alkaline, pH 9–11, 45–55 °C, 3–5 min | Acid or ammonia attack; smut on brass | Dedicated copper cycle, separate from the steel line |
| Rinse | Two stages, final on demineralised water | Salts dry on the part and block adhesion | Rinse water checked before each batch |
| Oxide removal | Mechanical key, or 5–10% sulphuric acid for 30–60 s | Over-pickling etches grain and leaves smut | Pickle only where the oxide is heavy; rinse immediately |
| Dry | Warm air or oven at 60–80 °C | Water spots and re-oxidation | Dry before the part cools below 40 °C |
| Coating window | Same shift, ideally within 4 h | Oxide regrowth under the film | Parts cleaned but not coated by end of shift are re-cleaned |
| Powder | Polyester 60–80 µm; epoxy 100–140 µm for insulation | Thin edges, thick flats | DFT gauge on copper; eddy-current or magnetic depending on substrate |
| Cure | Per supplier schedule, 180–200 °C metal | Overbake embrittles the film on a high-expansion substrate | Oven profile recorded per batch |
| Adhesion | Cross-cut ISO 2409, class 0–1 | Passes on the flat, fails at the edge | Cut across a formed edge, not just a flat face |
| Corrosion | ISO 9227, assessed by scribe creep | Red-rust criteria do not apply | Assess to ISO 4628 for blistering and creep |
Degreasing without etching
The degrease stage has to remove oil without removing metal. A non-etching alkaline cleaner at pH 9–11, held at 45–55 °C for 3–5 minutes, lifts drawing oil and handling soil while leaving the surface grain intact. Silicate-containing cleaners are best avoided on copper because silicate residues are hard to rinse and show up later as a haze under a thin film.
Temperature is a compromise. Hotter cleaner works faster and oxidises the surface faster once the part leaves the bath; 45–55 °C is the range where the oil comes off and the oxide does not run away. The same logic applies to rinse: warm demineralised water, then a quick dry, then the booth.
Removing the oxide without wrecking the surface
Where the oxide is light and uniform, a mechanical key is preferable: a fine alumina brush or a light glass-bead blast at low pressure. Never use steel shot on copper. Steel media embed fragments in the soft surface, and an embedded steel particle under a coating is a corrosion cell that shows as a rust spot on a copper part.
Where the oxide is heavy — after laser cutting, or after a part has sat in a humid store — a short pickle in 5–10% sulphuric acid at room temperature, 30–60 seconds, removes it. The risks are over-pickling, which etches the grain boundary and leaves a smut that has to be brushed off, and drag-out, which carries acid into the rinse. Brass pickles faster than pure copper and dezincifies if left too long, so the time has to be set per alloy rather than per part family.
The window
The interval between the final rinse and the powder booth is a process parameter, not a convenience. On a dry day a cleaned copper part begins to tarnish within 30–60 minutes; in a monsoon shop it is faster. Four hours is a workable limit for a clean, dry, indoor transfer, and parts that miss it are re-cleaned rather than coated.
That rule is easy to write and hard to run, because cleaning is a batch operation and coating is a queue. The practical fix is to clean in the quantity the booth can coat in the same shift, and to keep cleaned parts in a dry cabinet rather than on an open trolley.
Does copper need a conversion coating?
No — and that is the part of the steel process that has to be dropped rather than adapted. Copper gets no benefit from a phosphate film, and the chemistry does not form one properly.
What copper can take is a thin adhesion promoter: a silane or a proprietary copper-compatible passivation, applied at the end of the rinse and dried before coating. These are not conversion coatings in the phosphating sense; they are coupling layers a few tens of nanometres thick, and their value depends on the coating system and the service environment. Where the specification allows one, it buys tolerance to a few hours of storage. Where the customer's specification forbids it, the answer is a tighter cleaning-to-coating window, not a substitute chemistry.
For parts that will be soldered or welded after coating, the joint area is masked and left bare, and the bare area needs its own protection — usually a temporary tarnish inhibitor rather than a coating, because a film that has to be removed before soldering defeats the purpose.
Adhesion, cure and the expansion problem
Copper's thermal expansion is the number that catches out a coating that passed every bench test. At 16.5 µm/m·K against 12.0 for steel, the difference is 4.5 µm/m·K. Over a 180 K cure cycle and a 500 mm panel, the copper grows 500 × 4.5 × 10⁻⁶ × 180 = 0.405 mm more than the steel it is fixed to.
If the copper panel is rigidly constrained — riveted at close centres, or bonded to a steel frame along its full length — that 0.4 mm has to go somewhere. It goes into the film as shear at the edges, and a rigid, highly crosslinked coating will craze there after a few hundred cycles. Three fixes work: specify a flexible powder, reduce the constraint by using slotted fixings, or keep the coating off the joint face and let the parts move.
Cure is a metal-temperature question as it is on any substrate, with one copper-specific twist: copper conducts heat at about 385 W/m·K, roughly seven times steel, so a copper part reaches oven temperature quickly and a thin copper panel can be overbaked while a thick steel bracket on the same rack is still climbing. Rack copper separately, or accept a longer cycle at a lower set point.
Thickness measurement is the other practical difference. A magnetic DFT gauge does not read on copper; an eddy-current gauge does, calibrated on a copper shim of the same alloy. Reading a coating thickness with the wrong gauge is worse than not reading it, because the number looks plausible.
What salt spray tells you on copper
ISO 9227 neutral salt spray is a comparison test, and on copper the failure signature is different from steel. Copper does not red-rust; its corrosion products are green, white or brown, and a scribe that creeps will show a green stain rather than the orange bleed a steel part gives.
That matters for acceptance. A specification written for steel — "no red rust at 500 h" — will pass a copper part that has lost adhesion along every scribe, because there is no red rust to find. Write the requirement as scribe creep in millimetres and blistering to ISO 4628, with the assessment area and the magnification stated.
For interior parts, a condensation test is often more relevant than salt spray, because the service is humidity and temperature cycling rather than chloride. Either way, the method has to be named with its duration and its acceptance limits, or the laboratory number cannot be turned into a purchase decision.
Where Ganesh coats copper parts
Copper and brass components are laser cut on the 12 kW fibre machine, formed on the 200 T, 5-axis brake where the alloy and thickness allow, and finished through the seven-tank pretreatment line on a copper-specific cycle: alkaline degrease, two rinses, oxide removal where needed, demineralised rinse and dry. Coating runs in the 18 ft × 14 ft powder booth with the 16 ft × 16 ft blast chamber available for heavier oxide or weld discolouration.
The plant holds ISO 9001:2015 and ISO 14001:2015, and an effluent treatment plant handles the chemical side of the line, which matters more on copper than on steel because the rinse water carries dissolved metal. DFT and cross-cut adhesion are recorded per batch, and the copper route is kept separate from the steel phosphating stages.
The paint shop capability page lists the full sequence, and copper and brass fabrication covers cutting, forming and finishing for the alloys. The laser side of the same parts is covered in laser cutting copper and brass, and the general cleaning principles in degreasing before powder coating. Coating cost per part, including masking and batch size, is modelled in the powder coating cost estimator, and the surface treatment hub collects the substrate-specific guides.
What to put on the coating drawing
- Alloy and temper, because brass and pure copper pickle and form differently.
- Surface condition on arrival — oiled, tarnished, laser cut or previously plated.
- Cleaning restriction, if the customer's process names one, plus a note that the part must be water-break free before coating.
- Oxide removal method, mechanical or pickle, and the maximum pickle time per alloy.
- Coating system — polyester or epoxy, colour, gloss and DFT range.
- Masked areas: contact faces, earthing points, solder joints and any thread.
- Cure limit, especially where the assembly includes a component that cannot take 200 °C.
- Adhesion method and class, cross-cut to ISO 2409 at class 0–1.
- Corrosion test and acceptance, with scribe creep and blistering limits rather than red rust.
- Packing, because a coated copper part that rubs in transit arrives with a shiny edge and a failed film.
When powder on copper is wrong
Powder is a good finish for a copper cover, a bracket or a housing that has to match a colour and survive weather. It is the wrong finish in four situations.
Busbar insulation needs a dielectric coating with a defined breakdown strength and a holiday test, not a decorative powder. A 60 µm film might hold 2–3 kV, but a pinhole at a laser-cut edge holds nothing, and the test is what proves it. That is an electrical insulation specification, and it belongs with a coater who tests for it.
Soldered or welded joints have to be masked, and a masked area on copper tarnishes faster than the same area on steel, because there is no phosphate film to hold a temporary protectant in place. Specify a temporary tarnish inhibitor or a plated finish.
Marine and chemical service usually needs a different substrate or a plating system. Copper-nickel alloys resist seawater; coated pure copper under a damaged film does not.
High-thermal-cycle assemblies rigidly constrain the copper against steel, which is the expansion case above. A flexible coating and slotted fixings are the minimum; a different joint design is better.
Send the alloy, the surface condition and the service environment with the drawing, and the quotation will state the cleaning cycle, the coating window and the test method it assumes.
Frequently asked questions
Can you powder coat copper?
Yes, with preparation that suits the metal. Clean to a water-break-free surface, remove the oxide mechanically or with a short dilute acid pickle, coat within a few hours, and cure at the powder supplier's schedule. The film adheres well when the oxide is gone; it lifts along edges when it is not.
Why does powder coating peel off copper?
Because the weak layer is copper oxide, not the powder. Copper oxidises within minutes of cleaning and grows a thicker oxide during the oven ramp, so a part cleaned on Friday and coated on Monday has a 10–20 nm oxide under the film. Use a mechanical key or a short pickle, and coat the same day.
Can copper parts go through the same pretreatment line as steel?
Not through a zinc phosphate stage. Copper ions contaminate the bath and degrade the coating on subsequent steel parts, and copper itself does not form a useful phosphate film. Clean and rinse copper on a separate cycle, and keep the tanks dedicated.
What cleaner is safe for copper and brass?
A mildly alkaline, non-etching cleaner at pH 9–11 and 45–55 °C, silicate free, with a demineralised final rinse. Avoid acidic degreasers, ammoniacal cleaners and strong chelators, and never use an ammonia-based cleaner on stressed brass because it can trigger stress-corrosion cracking.
How is adhesion tested on a coated copper part?
Cross-cut to ISO 2409, class 0–1, plus a bend or impact test on a coupon from the same batch. Salt spray to ISO 9227 is also useful, but assess it by scribe creep and blistering to ISO 4628 rather than by red rust, which copper does not produce.
Sources and further reading
- Copper Development Association — cleaning and surface preparation of copper and copper alloys · reviewed 3 Nov 2026
- ISO 9227:2022 — Corrosion tests in artificial atmospheres: salt spray · reviewed 3 Nov 2026
- ISO 2409:2020 — Paints and varnishes: cross-cut test · reviewed 3 Nov 2026
Sending copper parts for coating?
Send the drawing with the alloy, temper, surface condition and finish. We will confirm the cleaning route, the coating window and the film build before quoting.
Request a coating quote