The surface finish of a PCB might not be the first design aspect a new layout addresses, but it should be among the top contenders.

Where do you usually begin a design? Mechanical design aspects would lead one to select rigid, flex or bend-flex. Operating temperature would dictate the required material is standard FR-4, a high Tg version, a polyimide, PTFE, or ceramic. Microwave requirements might lead to a low dielectric constant and dissipation factor resin. Heat dissipation might require aluminum or copper core boards. But what is required for the top surface where soldering and wire bonding takes place? The area where oxidation and metal migration occurs: have you addressed those concerns through careful selection of surface finish?

The oldest used finishes are SnPb solder (easily soldered at moderate temperatures but lead-containing) and OSP (very flat SMT pads but short shelf life). These were the mainstays of the industry, and processing equipment, solder pastes, fluxes, solder masks, and even stencil inks have been designed to handle those processing profiles.

The move to RoHS in Europe and later in the US (except for some mil spec and medical applications) resulted in development of lead-free (SAC) solders that are tin-based, with additions of silver and copper and with traces of bismuth, indium, zinc or antimony. These pass government RoHS regulations and after process development can minimize growth of tin whiskers. Initially, the higher soldering liquidus temperature led engineers to specify high glass transition composite substrates (170°-180°C Tg) but careful attention to reflow profiles allowed a return to the less-expensive standard Tg 135°C FR-4 materials. Since the global industry has moved further toward RoHS, many circuit board manufacturing plants have standardized on Pb-free HAL and no longer price that surface at a premium. Indeed, some have scrapped HASL entirely and subcontract that finish, resulting in a cost premium for leaded solder.

Another option, albeit more expensive than Pb-free HAL, is ENIG (electroless nickel/immersion gold). This surface, like OSP, renders the pads very flat and offers the additional advantage of longer shelf life and the ability to wire bond. There is a low but measurable incidence of gold-to-nickel adhesion failure. This is called “black nickel” or “black pad” (see photo, below right) because, not surprisingly, the surface of the nickel turns black. The immersion gold on the pad surface looks normal and solder wets to the gold quite easily. However, with stress or even vibration the components could easily pop off without warning, the result of a brittle fracture at the interface. There are theories about phosphorous levels in the nickel plating solutions being off-spec and causing intermetallics during the gold plating. This problem is rare but does occur.

The most recent advance in surface preparation is ENEPIG, or electroless nickel/electroless 

palladium/immersion gold. Typical thicknesses are 150-200µin of nickel followed by 5µin palladium and then 1 to 3µin of gold. There is a cost premium over ENIG but the advantages could far outweigh

 the increase.

In ENEPIG the palladium interlayer acts to protect the nickel layer during the gold plating process but dissolves into the solder during reflow, yielding an oxide-free activated site for an optimum bond to the components. The gold and palladium layers are inert and do not oxidize, resulting in excellent shelf life. Once the gold and palladium are dissolved during the soldering process, the remaining layer of nickel prevents copper leaching into the solder.

For all of the advantages inherent in each surface treatment, what are the associated costs? The cost increase related to the precious metals is directly related to the surface area being plated, but given a typical mid-range circuit board a cost comparison would in today’s metals market yield:

The cost increase with ENEPIG compared to Pb-free HAL is significant but if it does the job, it is well worth the cost. On the other hand, the 28% premium for ENEPIG or even the modest 14% increase with ENIG could be money saved if the components do not demand the flattest of pads. Consider the processing parameters of the BoM early in the design process as the surface requirements will significantly affect the overall board cost.

Robert Sergio Simon is a veteran of technology development and marketing, having worked R&D and marketing for electronics, polymer, and metal companies, including Bayer AG and Battelle Memorial Institute, before founding USTEK Inc.; This email address is being protected from spambots. You need JavaScript enabled to view it..