Gold Thickness: 5 Checks for ENIG Finish

Gold thickness on an electroless nickel immersion gold finish is a compromise rather than a target. Enough gold protects the nickel and keeps the surface solderable through storage, while too much gold makes the joint brittle and adds cost that no customer wants to pay for.

The process is also unusual, because the gold is deposited by a displacement reaction that stops by itself once the surface is covered. That self limiting behaviour is convenient, and it is also the reason the deposit is so sensitive to the state of the nickel underneath it.

What Gold Thickness Has to Achieve

The layer has two jobs. It protects the nickel from oxidation so that the board keeps its solderability, and it provides a surface that the solder can wet quickly during reflow. Both jobs are complete at a thickness far below the figures sometimes quoted in specifications.

The industry range for immersion gold is narrow, and the upper limit exists because gold dissolves into the alloy and forms brittle intermetallics. A joint with excessive gold can look perfect and fail a drop or a thermal cycle test months later.

Immersion Gold and the Nickel Beneath

The displacement reaction needs clean, active nickel. Where the nickel has oxidised, or where the bath has been contaminated, the gold deposits thinly in some areas and not at all in others, and the result is a patchy surface that reads low in one place and acceptable in another.

The nickel layer also carries the solder joint. Its thickness and its phosphorus content decide how the joint forms, and a nickel deposit that is too thin will not stop the copper from diffusing into the gold. Our board finish notes compare the alternatives to this finish.

Measuring with X-Ray Fluorescence

X-ray fluorescence is the standard production measurement, and it is fast, non destructive and reasonably accurate on gold. The method needs a calibration standard matched to the finish, and it needs to be checked on a known sample rather than trusted from the display alone.

The measurement spot is small, so the reading describes the point under the beam rather than the panel. Several points across a panel, including a pad in the centre and one at the edge, describe the distribution far better than a single reading taken at a convenient place.

Thin Gold and the Blackening Mechanism

Where the gold is thin, the nickel underneath is exposed to the plating bath and to the air, and it corrodes along the grain boundaries. That corrosion produces the dark, glassy surface that gives the defect its name, and the joint that forms on it is weak and often brittle.

The mechanism is a process problem rather than a material one, and it usually points at the nickel bath or the gold bath condition rather than at the thickness alone. A panel with blackened pads in a local area is telling you which part of the tank or the rack was at fault.

Thick Gold and the Cost Argument

Every extra tenth of a micron of gold is paid for twice: once in the metal and once in the time the panel spends in the bath. Since the protection is complete well below the upper limit, a thicker deposit adds cost without adding performance.

Where a customer asks for a higher figure, the reason is usually a specification carried over from a different finish or a wire bonding requirement. The distinction matters, because a bondable gold is a different process from a solderable one and cannot be substituted by thickness alone.

Solderability and the Storage Window

The finish has to survive the storage the product will see, and the gold protects the nickel on that timescale. A board stored for a year in a warm, humid place will still be solderable if the finish and the packaging are correct, while a board stored in a paper sleeve in the same room may not be.

Packaging, handling and shelf life therefore belong with the finish specification rather than the warehouse, and our shelf life notes describe how the window is controlled for the whole assembly.

Bath Control and the Gold Concentration

The immersion gold bath depletes its gold as panels are processed, and inhibits its own reaction as it loads with nickel. Both effects make the deposit thinner with use, so the concentration and the contamination level have to be analysed and dosed rather than topped up on a hunch.

Temperature and immersion time are the other two controls, and they act in opposite directions to the bath age. A bath that is run until the deposit falls below the limit and then replaced on a planned stop gives a far more consistent result than one that is pushed until a defect appears.

Panel Features That Read Low

Small pads and isolated features plate more slowly than large ones, because the local supply of gold is quickly consumed. The measurement on a fine pitch pad can therefore read below the minimum while the large test pad on the same panel is comfortably inside the range.

This is a design and a measurement issue rather than a process failure, and the acceptance criterion should state which features are measured. Our plating uniformity notes describe the same geometry effect in copper plating.

Acceptance Criteria and Reports

The specification should state the minimum gold thickness, the nickel thickness, the measurement method and the features to be measured. Without all four items, two laboratories can test the same panel and produce answers that cannot be compared.

A report should also state the sampling, because a single point on one panel says very little about a batch. Where the customer requires a certificate, the data behind it should exist, and our anode bag notes show the kind of plating line records that support one.

Records for a Finish Claim

A claim against a finish needs the panel identity, the bath analysis, the thickness readings and the soldering conditions used by the assembler. With those four items the discussion can be about the process rather than about the gauges.

Where the failure is a blackened joint, the section of the failed board is the decisive evidence. The IPC finish standards define the acceptance criteria and the tests that should be used to establish whether the deposit met them.

FAQ

Is a thicker gold layer always more reliable? No. Beyond the recommended range the gold embrittles the joint and can reduce its fatigue life. The requirement is a window rather than a minimum, and the upper limit should be stated as clearly as the lower one.

Why do some pads measure low on the same panel? Current distribution, feature size and the local condition of the bath all affect the deposit, and small isolated features are the first to read low. The acceptance criterion should say which pads are measured so the result is meaningful.

Does the gold thickness change during storage? It does not grow or shrink, but the surface can become contaminated and the nickel beneath can corrode if the gold is porous or thin. Storage is therefore a factor in the finish choice even though the deposit itself is stable.