PCB Plating Thickness Guide: Copper, Nickel and Gold

Three Metals, Three Jobs

Plating thickness is one of the few numbers on a fabrication drawing that reaches into electrical performance, mechanical life and cost at the same time. Copper carries current and holds the barrel together, nickel stops copper from diffusing into gold, and gold keeps the contact surface solderable and corrosion free. Each has its own working range, and going above it costs money without buying reliability.

This guide collects the ranges that appear on real drawings, explains where each number comes from, and shows how the layers are checked in production.

Electroplating and Electroless Plating

Two deposition methods share the shop floor. Electroplating drives metal out of solution with a current, which is how copper is put into drilled holes and how hard gold is applied to edge connectors. Electroless plating uses a chemical reduction instead, with no current involved, which is how the thin copper seed in the hole and the nickel gold stack of an ENIG finish are formed.

Electroless copper is the enabling step: it covers the drilled resin and glass so that the subsequent electroplated copper has something continuous to build on. If that seed is thin or broken, the barrel fails later at temperature cycling, long after the board has shipped.

Copper: Foil Plus Plating

The copper figure people quote is a sum, not a single deposition. Start with the foil the laminate arrives with, add the copper plated during processing, and the total is what carries the current and what the impedance calculation should use.

• Inner layer foil: 0.5 oz to 2 oz (roughly 17 to 70 micron).
• Outer layer foil: 1 oz to 3 oz before plating, and effectively more afterwards.
• Plated barrel: 20 to 25 micron is the usual specification, with 25 micron as the minimum for many automotive, medical and industrial programs.

The difference between foil weight and plated thickness confuses drawings more often than anything else. One ounce of copper is about 35 micron of foil. A hole wall that ends up at 25 micron of plated copper plus the foil lands around 90 micron at the surface land, which is why a drawing asking for 1 oz finished copper and one asking for 25 micron barrel copper describe very different processes.

Barrel thickness drives three things at once: current carrying capacity, resistance to thermal cycling, and the pressure the drill and the press put on the hole. Thin barrels crack during reflow; thick barrels are harder to drill cleanly and cost more to plate. For designs that carry serious current, the answer is usually more copper on the outer layers rather than more in the holes, which is exactly the trade discussed in heavy copper PCB fabrication.

Nickel and Gold in an ENIG Finish

The electroless nickel immersion gold stack is the default finish for dense SMT assemblies, because it gives a flat surface over fine pitch pads. The nickel is not decorative. It is a diffusion barrier that keeps copper from migrating into the gold, and it provides the hardness that stops a pad from deforming under a probe or a connector.

• Nickel: 3 to 6 micron, with 4 to 5 micron a common target.
• Gold on ENIG: 0.05 to 0.1 micron. Thicker gold is not better here, because a thick immersion layer becomes brittle and can cause black pad or solder embrittlement.
• Hard gold on edge connectors: 0.5 to 2 micron over 3 to 6 micron of nickel, plated with cobalt or a similar hardener for wear resistance.

That last range is worth pausing on. A gold finger in a card edge slot may see hundreds of insertions, and the wear life scales directly with the gold thickness. Buying 0.5 micron when the application needs 1.5 micron saves money once and costs a field failure later. Beyond ENIG there are other finishes with their own thickness logic, and the comparison in immersion silver PCB processing is a useful reference point when choosing between them.

How the Layers Are Measured

Two techniques dominate incoming and final inspection.

X ray fluorescence measures the mass per unit area of each metal in a stack and converts it to thickness. It is fast, non destructive, and works well on gold and nickel over a flat pad. It is less precise on very thin layers and on curved surfaces, and it cannot see inside a hole.

Microsectioning answers the questions XRF cannot. A coupon is potted, ground, polished and viewed under a microscope, which reveals the actual barrel thickness at the middle and the ends of the hole, the plating distribution between the surface and the centre, and any voids or cracks in the copper. This is the method behind the 20 micron minimum, and it is the one named in most acceptance criteria.

Production control usually combines both: XRF on the panel for a quick reading, microsection on the coupon for the record. That pairing is a normal part of quality management on any program with reliability requirements rather than a special request. Both sets of numbers should appear in the first article report.

What Moves the Thickness in Production

Plating is not uniform by nature, so thickness is a distribution rather than a value.

• Current density. Higher density plates faster but concentrates metal at the hole entry, leaving the centre thin. This is the classic dog bone shape in a microsection.
• Bath chemistry. Additive packages are what allow a bath to throw copper into the middle of a deep hole instead of plating the surface only. Bath ageing, contamination and temperature all shift the result.
• Hole geometry. Aspect ratio is the controlling variable. A 0.3 mm hole through a 3 mm board is a very different plating problem from a 0.3 mm hole through a 0.6 mm board.
• Design density. Isolated holes plate differently from a field of dense vias, because the local current demand differs across the panel.
• Equipment. Pulse plating and well controlled agitation narrow the spread between the hole entry and the centre, which is why the same drawing can pass at one shop and fail at another.

The practical consequence for a designer is that a specification should state the minimum at the worst point of the hole, not an average. Any note that says 25 micron without saying where invites a measurement argument later.

Where Thickness Meets Performance

Current capacity. A 1 mm wide outer trace at 1 oz carries roughly 1 A with a moderate temperature rise. Doubling the copper halves the resistance and the heating, which is why power stages, motor drives and LED boards specify 2 oz and above.

Impedance. Plated copper changes the trace cross section after lamination. Because a controlled impedance coupon is built on the same panel, the coupon sees the same plating and the impedance is measured in the finished state, not in the design tool. This is one of the reasons the coupon result and the simulated value can differ slightly, as described in TDR impedance testing.

Thermal cycling. A barrel with 12 micron of copper survives a few hundred cycles. A barrel with 25 micron survives thousands, because the copper can stretch with the resin instead of cracking. Automotive and medical programs pay for the difference.

Solderability. The gold or silver layer protects the pad until the paste is printed. A pad with an incomplete finish passivates and produces a solder defect that looks like a paste problem, not a plating problem. Reflows behave differently on a contaminated surface, which is worth remembering the next time a board is blamed during PCB assembly.

Cost Impact

Only some of the plating range is charged for.

• Standard copper plating: included in the price of the board.
• Heavy copper plating: typically adds 20 to 120 US dollars per batch depending on the target weight and the number of steps.
• ENIG finish: around 0.03 to 0.12 US dollars per square inch of board area.
• Hard gold on connectors: 80 to 400 US dollars per batch, driven mostly by the thickness and the plated area.

A useful rule is to specify the minimum that meets the reliability requirement and no more. Adding gold that the application never wears, or copper that the net never uses, raises the price of every unit for the life of the product.

FAQ

What is the standard via copper thickness? 20 to 25 micron in the barrel is the common requirement, with 25 micron used for higher reliability classes.

How thick is ENIG? Nickel 3 to 6 micron and gold 0.05 to 0.1 micron, measured over the nickel.

Can the plating be verified on my boards? Ask for XRF readings on the panel and a microsection photograph of the coupon in the first article report.

Is thicker gold always better on a connector? Thicker hard gold extends wear life, but thicker immersion gold on ENIG pads is a defect, not an upgrade.

Summary

Plating thickness is a specification with a purpose behind every number. Twenty to twenty five micron of copper in the barrel keeps the hole electrically and mechanically sound, 3 to 6 micron of nickel blocks diffusion under the finish, and 0.05 to 0.1 micron of gold protects a pad through one assembly while 0.5 to 2 micron of hard gold protects a connector through years of insertions. Specify the minimum, state where it is measured, and confirm it with both XRF and a microsection rather than assuming the drawing was followed.