ENIG Gold Thickness Control And Black Pad Risk
Electroless nickel immersion gold is one of the most widely used finishes on a printed circuit board, and the part of it that draws the most argument is the thickness of the gold. The complaint is always that there is not enough of it, and the answer is almost always that there is enough for the purpose it was specified for, provided the nickel underneath is sound. Understanding why the gold layer is deliberately thin, and what makes it thicker or thinner in practice, removes most of the confusion.
This article explains what the two layers do, why the gold is measured in fractions of a micron, how the process can go wrong, and how the finish is verified on the shop floor and in the laboratory.
What The Finish Is Made Of
ENIG is a two layer system. A layer of nickel is deposited on the copper by an autocatalytic chemical reaction, and a much thinner layer of gold is then deposited on the nickel by a displacement reaction in which the gold in solution exchanges places with a small amount of the nickel. The nickel is the barrier that stops copper and tin from meeting during soldering, and the gold is a protective layer that keeps the nickel from oxidising before the board is used.
Both layers are deposited without current, which is what allows the finish to cover isolated pads, fine traces and the inside of small vias evenly, including edges that an electroplated finish would leave thinner. That coverage is the main reason the finish is chosen, and it is also the reason the chemistry is sensitive to the surface it is given.
Why The Gold Is Deliberately Thin
Gold thickness in a standard ENIG finish is normally quoted between 0.05 and 0.1 microns, with a minimum of 0.04 microns for a class 2 product and a value near 0.075 microns appearing on many drawings. The number is small because the layer only has to be continuous, and because the displacement reaction that deposits it slows down as the layer grows. Once the surface is covered with gold, the exchange with the nickel underneath has nothing left to react with, so the process stops itself.
Pushing the gold thicker is not simply a matter of leaving the boards in the bath longer. A prolonged immersion continues the attack on the nickel, which deepens the corrosion of the grain boundaries and increases the risk of the condition known as black pad. A heavier gold also raises the cost sharply, dissolves more nickel into the bath, and makes the solder joint more brittle, because gold dissolves readily into molten tin and forms intermetallic compounds.
What The Nickel Layer Contributes
The nickel is typically specified between 3 and 6 microns, and its quality controls almost every problem attributed to the gold. It has to be dense, adherent and free of the excess phosphorus that comes from the reducing agent in the bath. A nickel layer with 7 to 10 percent phosphorus is amorphous and acts as a good barrier; a layer with a higher phosphorus content, or one whose chemistry has drifted, becomes more prone to corrosion at the grain boundaries.
The nickel also has to be deposited on copper that has been properly activated. If the copper surface was passivated or contaminated, the nickel will start in patches and the finish will later show as a discoloured pad or a solder joint that fails at low load. Reviewing the gold thickness without looking at the nickel is therefore a diagnosis of the wrong layer.
How The Process Goes Wrong
Two failures account for most complaints. The first is a thin or discontinuous gold layer, which leaves nickel exposed and produces pads that oxidise in storage and solder poorly. It usually traces back to a bath that is low in gold concentration, to a temperature that has fallen, to an immersion time that is too short, or to a very high surface area of copper in the load, all of which reduce the driving force of the displacement reaction.
The second is black pad, in which the nickel under the gold corrodes along its grain boundaries, leaving a surface that wets slowly and a joint that fractures in a brittle way. It is associated with a nickel bath that is over active, with a gold bath whose pH or temperature has drifted, and with too long an exposure. The failure is not visible on the surface, which is why a section or a solderability test is needed to confirm it.
Measuring The Thickness
The routine measurement is made by X-ray fluorescence, which reports the thickness of the gold and of the nickel beneath it without damaging the board. The method is fast and can be used on the line, but the reading depends on the area measured, on the calibration standard and on the composition of the alloy under test, so a figure that is compared between two machines has to be produced with the same set-up on both.
Where the result matters legally or where a dispute has arisen, a microsection is used instead. The board is cut through a pad and the layers are measured at magnification under a calibrated scale. The section also reveals the shape of the nickel surface and the depth to which it has been attacked, which is exactly the information the XRF reading cannot provide. Coupon panels plated with the production load are the usual way of keeping both methods aligned.
Choosing A Thickness For The Application
The requirement follows the joint, not the finish. A board that will only be soldered needs enough gold to protect the nickel, and no more; a heavier layer adds cost and embrittlement risk without improving the joint. A board that will carry wire bonds is a different case, since the bonding process needs a known, uniform and softer surface, and that demand is met by a separate gold specification applied over the same nickel barrier.
Boards that will be assembled more than once, or that will be stored for a long period before assembly, sit between the two. Extra thickness helps only if it is uniform, and an uneven layer is worse than a thin one, because it produces joints that behave differently across the same panel. Where the finish interacts with the assembly process, the interactions are described under PCB design and fabrication and manufacturable design guidelines.
Control On The Line
The bath is controlled by analysis rather than by counting boards. Gold concentration, pH, temperature, specific gravity and the level of dissolved nickel are measured on a schedule, and the load, measured as surface area per litre, is held inside a range that the chemistry can support. Racking matters too, since a panel that shields another panel produces a different result on the two sides of the same sheet.
The finished boards are sampled for thickness and for solderability, and the results are kept against the lot. When a customer reports a finish problem months later, that record is the only way to tell whether the boards were made inside the window or whether the complaint refers to a genuine excursion. Related assembly side effects of the same chemistry are covered under lead free versus leaded solder.
FAQ
Is more gold always better? No. Beyond the thickness needed to cover the nickel, extra gold raises cost, embrittles the joint and increases the attack on the nickel layer during deposition.
Can ENIG be reworked if the thickness is low? A board cannot be re-immersed to add gold, because the second pass attacks the nickel that is already there. Rework means stripping the finish and plating again, which is rarely economical on a finished panel.
How can black pad be confirmed? By a microsection through a failed joint, by a solderability test that shows slow wetting, or by examining the fracture surface of a joint that broke at low load and finding a nickel rich surface.
