As many readers know, the new century has not been kind for the North American printed circuit manufacturers. Although the first year saw record levels, the next three years were miserable. The industry had never experienced a three-year slump before—and certainly never one of such severe proportions. It was an unfortunate combination of slow orders (especially in telecommunications) and movement to China (roughly half of the production capacity in this country is lost). While business has been rebounding for the survivors, the board manufacturers are facing a constant battle with production changes.
Changes have been dramatic in the drilling (ever smaller holes), etching (much more sophisticated texturing of copper surfaces for enhanced solder mask adhesion) and thickness of the multi-layers required—including buried components and blind vias; laser-drilled blind vias, drilled holes are getting smaller, soldermask dams as low as 1.0 mil and thicker high count multilayer boards are more common. All these changes pose major challenges for the supplier and the board manufacturer.
The “final finish” used to be simple—either HASL (hot air solder leveling) or electrolytic nickel/gold. With the rapid increase in technological improvements, it became necessary to develop more sophisticated surface finishes. The main driver was to produce pads with better planarity, as well as finished pads for wire bonding and for press fit connectors. The surface finish is also called upon to serve as a contacting surface.
As the boards continue to shrink in size, the surface geography becomes more valuable, making bussing for electrolytic processes virtually impossible. This drove the industry to use the newer electroless processes being developed. There is also the perceived environmental benefit to eliminate the lead in the HASL process. Ridding the planet’s electronics of leaded solders has become the holy grail despite our own EPA’s studies that such lead doesn’t readily leach when in land fills.
“Final finish” implies one choice but this has not been the experience. While even recently the HASL process was still used on more than 70% of boards in North America, that won’t be allowed soon on products going to Europe and Japan, so this continent is already changing to be able to offer like products in all countries/markets. Many options have been brought to the market for coverage of the exposed copper surfaces. The global industry has tried various organic solderability preservatives (OSP’s), immersion tin, immersion silver, electroless nickel/immersion gold (with electroless gold as an optional next step) and a new-style gold direct on the copper. Each has benefits and potential weaknesses.
OSP’s come in thin and thicker versions. These are relatively inexpensive and easily applied but are limited in the number of heat cycles that can be obtained in subsequent assembly. Some OSP's require a nitrogen atmosphere at assembly. They are also not suitable for wire bonding or as a contacting surface.
Immersion tins have had some early success as a solderable surface but may have a limited future. The main weaknesses are the use of a carcinogenic ingredient (thiourea) and the evidence of occasional whiskering as well as intermetallic formation. Whiskers are especially a concern, with fine line and spaces where such whiskers could be knocked out of a hole during part insertion, thus raising the possibility of a subsequent electrical short. Copper/tin intermetallic forms during deposition and continues to grow thus limiting the useful shelf life of the stored parts. In this age of cost awareness, another factor that has limited this product’s growth is that it is almost as expensive as electroless nickel/immersion gold—without the additional benefits.
Immersion silver would seem to have a bright future. It is easy to apply to the boards, relatively inexpensive and usually performs well. Like the OSP’s, thin (2-5 uin) and thicker (8-12uin) deposit versions have been sold but the preference seems to be towards the thicker products. To prevent tarnishing, the processes have included an anti-tarnish as an ingredient within the silver bath or applied in a subsequent step. Current testing is looking for methods to provide complete coverage on the walls of through holes and into blind vias (holes of various sizes and depths that don’t completely penetrate multi-layer boards). Other process concerns are the possible inclusion of voids in the solder joint and better thickness uniformity per part. Some manufacturers have complained about issues with corrosion of the copper surface near holes. If severe enough, this could lead to shorts (thus failure of the board).
Electroless nickel/immersion gold (“ENIG”) has been growing steadily in use. It is the most expensive of the final finishes but offers the most benefits. This process also requires the most steps (cleaner/microetch, activator—usually a palladium catalyst,electroless nickel and immersion gold). Parts must be clean and have a smooth copper surface on which to build.
The electroless nickel is an auto-catalytic process that deposits nickel on the palladium catalyzed copper surface. The process requires continuous replenishment of the nickel ion and the reducing agent. Good process control (constituent concentration, temperature and pH) is the key to a consistent reproducible deposit. It is very important that the nickel be able to plate a surface with consistent phosphorus levels. Most prefer a middle range of 6-8% P, too low would easily corrode/too high makes subsequent soldering of parts more difficult.
Immersion golds are replacement chemistries. This means that they attach themselves to the nickel by replacing atoms of nickel with atoms of gold. The purpose of the immersion gold layer is to protect the nickel surface until such time as it is soldered to. The recommended gold thickness is 2-4 uin. As the purpose of the gold layer is to maintain the solderability of the nickel surface, it is necessary that it be thin (two to four microinches is preferred) and pore-free.
ENIG is by its nature a sophisticated chemical system that requires knowledgeable plating personnel and lab maintenance. “Dip and dunk” mentalities will only create defective parts.
This process has had various issues to deal with as it’s grown in use. Early versions have had problems with:
- background plating (speckled areas on the solder mask that was not intended to be plated), bridging or “foots” (extra- neous plating between lines thus causing shorts),
- incompatibility with solder masks,
- skip plating (where some pads aren’t plated with enig—though sometimes gold does adhere to the copper) often due to surface contamination on the copper or a static charge.
- the most notorious issue, black nickel (such that the nickel surface has become passivated or corroded, even if mostly still covered by the immersion gold layer). Black nickel is a more serious issue as it is hard to detect but can lead to part failure. It has been shown that there are several root causes: a non-uniformly catalyzed surface due to tin or soldermask residues, an improperly maintained nickel process; a poorly designed nickel chemistry that can’t provide a steady phosphorous content on the part; some board designs; and/or an aggressive immersion gold process. A phosphorous content in the nickel below 4% by weight will corrode during gold deposition and cause black nickel; over 11 or 12 % makes subsequent soldering difficult or impossible.
The most successful ENIG chemistries running today have had several design strengths: low palladium/no chloride activators for solder mask compatibility and to avoid background and skip plating; a mid-phos nickel that runs at lower temperatures, doesn’t need frequent dummy plating, holds its ability to perform even after repeated heatups and gives level plating around the shoulders of pads and lines; an immersion gold that has gentle chemistry and is self limiting in thickness.
To enhance the performance of the nickels, sophisticated controllers are being used to make additions and also provide history to the quality department. To maintain steady state in the nickel chemistry, the better controllers bleed in the ingredients in measured amounts versus major ads made periodically.
Newer combination immersion golds are being used where the first microinch is the traditional replacement/immersion type. The next microinch or two are attached through an autocatalytic procedure that allow gold to attach itself to gold in a non-replacing build-up. This gives rise to a non-corrosive bath that produces a pore-free gold surface. Direct gold over copper (also called DIG—“direct immersion gold”—although that’s somewhat of a misnomer as this is more than an immersion process) was developed specifically for parts where nickel could create RF interference. The planarity is exceptional—again dependent on the quality of the copper surface. To avoid the inherent problems of copper migration through the thin gold surface, it is necessary for these parts to go to final assembly within four months. Even that period is only possible with a combined replacement/ autocatalytic process. Normal immersion golds will have trouble properly attaching to the copper surface and providing the necessary pore-free layer.
One last option has been developed for more sophisticated boards intended for gold wire bonding. A neutral pH, autocatalytic electroless gold is in use that allows thicker gold deposits on top of ENIG without harming the solder mask. The more traditional, high pH electroless gold is great for non-solder masked parts.
Electroless palladiums have also been created but have not grown in use due to the higher, more volatile cost of palladium metal. The additional hardness benefits will provide some market use going forward as these chemistries are mastered and get accepted.
Constant change will continue to face all fields. The successful companies will be open to it while also, hopefully, making wise choices. Not all chemistries are created equal, and sophistication demands wiser choices in buying the process designed for the specific product’s end use.