ENEPIG Finish: Selection, Reliability and Solder Joints

An electroless nickel electroless palladium immersion gold finish, usually shortened to ENEPIG, was introduced for a specific problem: a board that has to be soldered and also wire bonded. The older gold finishes served one of those duties well and the other badly, and the palladium layer was added to keep the gold thin enough to solder without letting the nickel underneath oxidise. Selecting it is a decision about the assembly route as much as about the finish itself.

What the ENEPIG Stack Looks Like

The stack is deposited in three steps on the copper pad: a layer of electroless nickel, then a thin layer of electroless palladium, then a very thin layer of immersion gold. The nickel provides the barrier and the mechanical base, the palladium keeps the nickel from oxidising and slows the diffusion of gold, and the gold protects the palladium during storage.

Typical thicknesses are a few micrometres of nickel, a fraction of a micrometre of palladium and a gold layer measured in tens of nanometres. The gold is deliberately thin, because a thick gold layer dissolves into molten solder and forms brittle intermetallics that weaken the joint.

Where ENEPIG Sits Among the Finishes

Hot air levelling is cheap and robust, but its surface is uneven and unsuitable for fine features, while immersion silver and organic coatings are flat and inexpensive with a limited storage life and no wire bond capability. ENIG offers a flat surface and wire bonding, with the risk of the corrosion known as black pad.

ENEPIG replaces the direct gold on nickel interface with palladium on nickel, which removes the mechanism behind that corrosion and allows a thinner gold layer. The trade is cost: the palladium bath adds a process step, and the chemistry has to be controlled more tightly than a simple two layer finish. Where the assembly is mixed, with a connector soldered onto a board that also carries wire bonded devices, that extra control is easier to justify than it is on a board with a single process route.

Wire Bonding and Soldering on One Surface

Wire bonding needs a surface that will form a reliable intermetallic with gold or copper wire at a controlled temperature, and gold on palladium on nickel provides it. The bond has to be made on a surface that has not been contaminated by flux or by careless handling.

The same pad has to accept solder later. A thin gold layer dissolves quickly in the alloy and exposes the palladium, which wets readily with most fluxes. The result is a joint that forms at a normal profile with no special measures, provided the gold thickness stays inside the specified range. The joint strength is then set by the palladium and nickel below, and the intermetallic that grows during reflow is thinner than the one formed on a thick gold finish, which is the main reason the finish tolerates several reflow passes better.

Cross section of an ENEPIG pad showing the nickel palladium and gold layers

Thickness Control in the Nickel and Gold Layers

Nickel thickness determines the barrier performance and the mechanical strength of the pad. Too thin and the copper diffuses through during later thermal cycles, while too thick a deposit becomes stressed and can crack at a via or a fine trace.

Gold thickness is the variable most often blamed for field problems, because a layer that is too thin will not protect the surface during storage and one that is too thick will make the joint brittle. Both limits come from the assembly requirement, so the drawing should state a range rather than a minimum. The way that thickness is specified and verified is covered in the guide to plating thickness.

The Palladium Layer and Why It Exists

Palladium is the layer that separates the nickel from the gold, and it does two jobs. It prevents the nickel surface from oxidising between the plating steps, and it acts as a diffusion barrier that keeps the gold from migrating into the nickel and the nickel from reaching the surface.

Its thickness is a compromise. A layer that is too thin will not cover the nickel completely, allowing oxide to form in patches, while an unnecessarily thick layer adds cost and can change the wetting behaviour during soldering. X-ray fluorescence is the usual measurement, and the instrument reports the gold and palladium layers together unless the software separates them, which is a common source of disagreement over a shipment.

Wire bonded pad beside a soldered pad on an ENEPIG finished board

Black Pad and Related Failure Modes

Black pad is the classic failure of the older ENIG finish, in which the nickel surface is corroded during the immersion gold step and the resulting joint fractures at the interface. Its signature is a dark, brittle pad and a joint that fails mechanically rather than electrically.

Other modes are shared with any plated finish. A nickel layer that is too thin lets copper diffuse to the surface, residues left by an exhausted bath cause poor wetting, and a contaminated palladium bath produces a surface that solders unevenly across the pad. The defects that follow from these conditions are collected in the guide to solder defects and board failures, and they are worth reading as a set, because poor wetting on one pad and a fracture on another can share the same plating origin.

Solder Joint Behaviour and Reliability Testing

Joints on ENEPIG are generally comparable with ENIG in thermal cycling, with the advantage of a thinner gold layer and a more stable interface. When a failure does occur it is usually at the nickel interface, so cross sections and pull tests tell the engineer more than a simple continuity check.

For wire bonded assemblies the bond pull test and the ball shear test are the standard measurements, and they should be performed before and after the environmental exposure the product requires. The methods and their limits are set out in the guide to solderability testing.

Process Control at the Fabricator

The plating line is where the finish is made or lost. Bath concentration, temperature, pH, dwell time and the condition of the palladium and gold baths all affect the result, and the gold bath in particular becomes contaminated as it is used, which changes the deposition rate.

Control means scheduled analysis, a record of the ampere hours or the square metres plated, and a coupon that is plated with every production panel. A coupon that is measured and archived turns a dispute about finish quality into a measurement. It also gives the assembly house something to compare against when a wetting problem appears, since the same panel carries both the production surface and a witness of the plating conditions that produced it.

Selection Checklist and Cost

A selection decision should be made against four questions: does the product need wire bonding, what storage life is required, how fine are the features, and what does the assembly process allow for gold thickness. If wire bonding is not required, a simpler finish will usually do the job at a lower cost.

Cost is not only the plating price. ENEPIG adds a process step and a stricter control burden, and it can reduce assembly yield if the gold thickness is wrong. The wider comparison of finishes is set out in the guide to HASL and other surface finishes.

FAQ

Is ENEPIG better than ENIG? For assemblies that need both soldering and wire bonding it is, because the palladium layer removes the mechanism behind black pad and allows a thinner gold layer. For a board that is only soldered, the extra cost is often hard to justify.

How thick should the gold be? Typically a few tens of nanometres, thin enough to dissolve into the joint without creating brittle intermetallics and thick enough to protect the surface in storage. The drawing should state a range rather than a minimum.

Can ENEPIG be reworked? It can, with the usual limit on the number of thermal cycles at the site. The risk on a reworked pad is mechanical rather than chemical, so the site should be inspected for lifting or damage before the replacement part is placed.

Leave A Comment