ENIG Surface Finish: Process, Thickness and Defects

A Thin Stack That Decides Assembly Yield

Bare copper oxidises, and an oxidised pad does not solder. The surface finish exists to keep the copper clean until the assembly line reaches it, and to do so without changing the shape of the pad. ENIG, electroless nickel immersion gold, is one of the most widely used answers because it deposits a very flat layer over fine geometry, which is exactly what dense boards with fine pitch devices require.

The finish is a two layer stack: electroless nickel as the soldering surface and the diffusion barrier, and a thin immersion gold layer on top to protect the nickel from oxidising before assembly.

The Two Layers and What Each Does

• Electroless nickel, usually 3 to 6 micron. It is the layer the solder actually bonds to. It provides mechanical strength, acts as a barrier that stops copper from diffusing into the surface, and gives the pad resistance to handling.
• Immersion gold, usually 0.05 to 0.1 micron. Its job is protection, not soldering. The gold dissolves into the solder as the joint forms, which exposes the nickel underneath and allows the joint to form properly. That is why a thicker gold layer is not an improvement on ENIG; it is a defect risk.

Compared with hot air levelling, ENIG produces a pad whose flatness is essentially the flatness of the copper underneath. Compared with an organic coating, it survives storage far longer and tolerates multiple reflow passes. The trade is cost and process difficulty.

The Process

• Cleaning and micro etching, which removes contamination and oxide and gives a consistent surface for the following steps.
• Activation, usually with a palladium catalyst, so that nickel will deposit on the copper.
• Electroless nickel plating, a chemical reduction that deposits a nickel alloy containing phosphorus. Bath chemistry, temperature and pH control determine the deposition rate and the phosphorus content.
• Immersion gold plating, a displacement reaction in which gold ions exchange with nickel atoms on the surface. It is self limiting, which is why the layer is thin.
• Rinsing and drying, with attention to what stays in the holes and under the mask.

Two process variables dominate the result. The nickel bath has to be maintained within a narrow window, because a drifting bath produces inconsistent phosphorus content and a plating rate that varies across the panel. And the immersion gold step has to be stopped before it attacks the nickel excessively, because over extended exposure the gold solution penetrates the nickel grain boundaries and starts the corrosion that produces the defect known as black pad.

Why It Is Chosen

• Flatness. The reason ENIG dominates dense assemblies. A flat pad gives a predictable paste deposit, which matters at fine pitch and under area array devices.
• Solderability over time. The finish survives months of storage and multiple reflow passes, which is why double sided assemblies with several thermal cycles use it.
• Electrical stability. The stable surface resistance suits fine geometry and high frequency work where a rough or oxidised surface would add loss.
• Wire bonding and contacting. Where the same board carries both soldered pads and contacts, a well controlled ENIG can serve both, though a dedicated contact finish is better where abrasion is involved.

The Characteristic Defects

Three failure modes account for most ENIG problems.

Black pad. The nickel surface corrodes along the grain boundaries during the immersion gold step, and the joint that forms there is brittle and weak. It appears as a dark pad, as solder joint fracture under thermal cycling, or as a failure that only shows up in the field. Control of the nickel bath, the gold immersion time and the pad geometry reduces the risk substantially.

Hyperactive nickel. The opposite failure, where the surface is too active and the nickel corrodes excessively or a thick gold layer forms beyond the self limiting point. It is a bath chemistry problem and it shows up as a rough, dark deposit rather than a flat one.

Incomplete gold coverage or a foreign deposit. Pads that are partly covered or contaminated will not wet properly, which reads on the assembly line as a solderability defect rather than a plating defect.

All three are controlled by the plater rather than by the designer. What the designer can do is avoid pad geometry that is difficult to plate uniformly, and specify the finish deliberately rather than by default. The alternative flat finishes, which are simpler to control, are compared under immersion silver PCB processing.

Thickness Specification and Measurement

The usual drawing values are 3 to 6 micron of nickel and 0.05 to 0.1 micron of gold. The gold figure is deliberately thin, and a customer who asks for a thicker gold layer on ENIG is asking for a worse joint, not a better one.

Thickness is verified by X-ray fluorescence, which is fast and non destructive, and its limitation is worth knowing: the measurement is sensitive to the alloy composition and can be affected by the nickel beneath, so the reading is a good process indicator rather than an absolute physical measurement. Where the value is critical, a microsection on the coupon provides corroboration. On a heavy copper board, the finish is applied over a large thermal mass and the plating time is longer, which changes the process window, so the specification and the measurement should both be confirmed with the fabricator.

Cost

ENIG is charged by plated area, and the price is informative because it shows how the finish interacts with the design.

• Typical adder: roughly 0.20 to 0.60 US dollars per board for small boards, and priced per square inch on larger ones.
• What drives it: the plated area, the number of pads, the presence of small holes that have to be plated consistently, and the gold thickness specification.
• What reduces it: keeping the finish only where it is needed, avoiding gold on areas that will be covered by solder mask or that do not need a flat surface.

The comparison to make is not ENIG against nothing, but ENIG against the alternatives. Hot air levelling is cheaper but gives an uneven surface that is unsuitable for fine pitch. An organic coating is cheap and flat but has a limited storage life and is not ideal for multiple reflow passes. Immersion silver and immersion tin sit in between. On a board with fine pitch devices, the calculated cost of an ENIG upgrade is small against the yield loss from a poor paste deposit. On a coarse pitch single sided board, it is money spent for nothing. Obtaining a comparison across two finishes for a single design is straightforward through a formal custom PCB pricing request, and the answer is often decisive.

When ENIG Is the Right Choice

• Fine pitch and area array devices, where paste volume consistency depends on pad flatness.
• Multiple reflow passes, as on a double sided assembly with fine parts on both sides.
• Long storage between fabrication and assembly, where a coating would tarnish.
• High density interconnect boards, where the pad geometry is small and the surface has to be exact. The surrounding stackup considerations are covered under HDI PCB fabrication.
• Products where soldering reliability is a safety requirement and the finish needs to be predictable across lots.

Where it is not needed: coarse pitch through hole boards, single sided products, and any design where a flat surface buys nothing and the additional process chemistry adds cost and risk without benefit.

Interaction With Assembly

Two points of interaction are worth planning for. Black pad risk is highest where the joints are small and the thermal cycles are severe, so a design that combines a fine pitch area array device with a high reliability requirement should treat the finish as a controlled process rather than a commodity. And the paste chemistry has to suit the finish; a solder paste developed for an organic coating may not wet ENIG as intended, which is worth confirming with the assembly house before the boards are ordered. Both points belong in the first article review and in the ongoing quality management of the program rather than being settled at the assembly line.

Questions to Ask a Supplier

Four questions establish whether the ENIG process is under control: what the nickel and gold thickness targets are and how they are measured, how the bath chemistry is monitored and how often it is changed, what the experience is on fine pitch boards, and how a black pad complaint would be investigated. The answers distinguish a shop with a controlled chemistry line from one that sends boards out to a job plater. Where the board also has to be assembled into a finished product, keeping the fabrication and the PCB assembly with one supplier removes an interface at the exact point where a finish problem first appears, and the wider process controls behind the finish are described under PCB manufacturing.

FAQ

What does ENIG stand for? Electroless nickel immersion gold, describing the two plating steps and the way each is deposited.

How thick are the layers? Nickel 3 to 6 micron and gold 0.05 to 0.1 micron. A thicker gold layer on ENIG is a defect risk rather than an upgrade.

What is black pad? A corrosion of the nickel surface that occurs along the grain boundaries, producing a brittle joint and a pad that can fail under thermal cycling. It is a process control issue.

Is ENIG suitable for gold fingers? No. A sliding contact needs electroplated hard gold of 10 to 50 microinch, not a thin immersion layer.

How much does ENIG add? Roughly 0.20 to 0.60 per board on small boards, or priced by area on larger ones.

Summary

ENIG is a two layer finish: 3 to 6 micron of electroless nickel, which is the surface the solder bonds to, and 0.05 to 0.1 micron of immersion gold, which protects the nickel until assembly. Its value is flatness and stability, which is why it dominates fine pitch and high density assembly, and its risks are the plating defects that result from an uncontrolled nickel bath or an over extended gold immersion step. Specify it where pad flatness and multiple reflow passes matter, and measure it with XRF on the panel and a microsection where the value is critical. The cost is small relative to the yield it protects on a dense board and wasted on a coarse one.