PCB surface finish options for lead-free manufacturing
Feb 23, 2005
Brian Toleno PhD, Application Engineering Team Leader, the electronics group of Henkel, Irvine, CA
Anybody affected by the forthcoming RoHS directive will know that it impacts not only upon the solder alloys that one is permitted to use, but also many other aspects of the electronics assembly process, including the selection and compatibility of board
Here we focus on some of the changes that are being made to surface finishes on printed circuit boards (PCBs) to make them compatible with lead-free, and look at their pros and cons.
For years, hot-air-solder leveling (HASL) was accepted as the most common and effective surface finish used to protect copper circuitry on PCBs. It wasn't until the move towards small fine pitch surface mount devices and area array packages - for which the non-planar nature of HASL coating causes problems - that a need for alternative surface finishes arose. With the advent of RoHS we now have other reasons to investigate alternative finishes, as the July 2006 deadline for compliance forces the industry to move away from Sn-Pb.
Suitable alternative finishes serve the same purpose as HASL; that of providing a solderable surface that prevents the underlying copper from oxidizing. There are several commonly-used alternative finishes, each of which has its own advantages and disadvantages that should be investigated before deciding on a particular surface finish. To this end, the IPC Alternative Finishes Task Group is studying the robustness of several surface finishes after exposure to a range of environmental conditions, and is expected to publish its findings during 2005. However, industry uptake of the most popular of the alternative finishes means that we already have some knowledge of their strengths and weaknesses, which we will discuss here.
Electroless nickel/immersion gold
When properly applied, the electroless nickel/immersion gold (ENIG) process produces a surface with good planarity and solderability, which can withstand multiple reflows. Since gold dissolves so readily into solder and does not tarnish or oxidize easily, it is an excellent choice for a surface finish. Unfortunately, the gold layer cannot be plated directly onto the copper, as the copper would diffuse into the gold, reach the surface and oxidize to form an unsolderable surface. For this reason, nickel is used as a barrier layer between the gold and the copper, as it plates easily onto copper and the gold plates easily onto nickel. Nickel is unsuitable for use as the outer layer, since it oxidizes quickly in air to form an unsolderable surface. IPC-2221 recommends an electroless nickel plating of 2.5-5.0 μm and an immersion gold layer of 0.08-0.23 μm. There are three commonly recognized failure modes with an ENIG surface finish: porous gold, gold embrittlement, and 'black pad'.
A porous gold condition is caused by gold atoms on the surface not forming a tight lattice, allowing the nickel to migrate to the surface and oxidize to an unsolderable surface. Using the immersion gold process, this condition rarely occurs and is usually not observed until the onset of solderability problems.
If sufficient gold is dissolved in copper, the abundance of the brittle AuSn4 intermetallic phase formed can cause solder joints to fail along the AuSn4 boundaries. This condition typically doesn't occur until the level of gold within the solder joint reaches more than four weight percent, therefore if the gold surface remains below 2.5 microinches (0.064 μm), it shouldn't become an issue. When obtaining leads with gold tinning, a solder dip process to dissolve some of the gold is recommended by J-STD-001.
The third condition produces weak solder joints that can crack and/or separate under minimal stress, an effect that is more pronounced in area array packages with their less compliant leads. It is termed 'black pad', due to its dark gray or black trademark left on the underlying nickel surface once the solder joint separates. The failure mechanism is less well understood than the other modes just described, but initial investigations into black pad have concentrated on plating bath chemistries, especially in relation to phosphorus content in the bath and residual phosphorus remaining after plating.
HADCO research into the black pad phenomenon
More recently, workers at PCB fabricator HADCO Santa Clara, Inc. reported a mechanism whereby the nickel underwent corrosion during the immersion gold (IG) process(1), terming this phenomenon 'IG hyperactivity'. Different stages of IG hyperactivity were shown to produce different amounts of corroded nickel, and as the level of corrosion increased, the strength of the nickel/tin intermetallic after soldering decreased. This decrease in Ni-Sn intermetallic strength directly corresponds to fractured solder joints and the observation of the black pad effect. The application of a low voltage through leads attached to QFP pads was also shown to be instrumental in multiplying the effects of the corrosion mechanism.
The Interconnect Technology Research Institute (ITRI) has also been investigating the black pad problem(2). Phase I concentrated on plating process variables, with no single process variable being identified as the sole cause of the black pad phenomenon, although certain chemistries were better than others. The findings also reported that the black pads observed were smaller pads connected electrically to a larger pad and/or a series of smaller pads. It was proposed that this produces a galvanic cell effect, similar to the application of a small voltage to drive IG hyperactivity as observed at HADCO.
Organic solder preservative
Another finish in widespread use is organic solder preservative, commonly known as OSP or Cu-OSP. As the name implies, the coating is organic in nature, and as it is transparent when applied, PCB tracks and pads boards retain their copper appearance. Early OSP coatings were blighted by poor shelf life and could only withstand one or two reflow cycles before their protective properties were degraded. However, the technology has undergone further development, and current OSPs are designed to withstand the Pb-free manufacturing process, and can be relied upon to protect the copper from oxidation even after multiple reflow.
Lower cost, but check your ICT
Because the coating is non-metallic in nature, it is both non-conductive and a lower cost process relative to some of the other alternative final finishes. While its non-conductive nature is generally regarded as beneficial, if the in-circuit test (ICT) points are OSP-coated, assemblers performing ICT must verify that the tester makes reliable electrical contact with the test pads after reflow, particularly if the PCB has been subjected to multiple reflow cycles.
Immersion tin and immersion silver
Two immersion finishes that have gained popularity recently are immersion tin (IT or ISn or white tin) and immersion silver (IAg). These finishes are not to be confused with the finishes of the same name offered several years ago; the current generations are much improved! Both finishes tend to be thinner than ENIG or traditional HASL and typically contain an organic component that retards the surface oxidation preventing one of the possible paths to solderability problems. The thinner surface can sometimes be an issue for press-fit connectors, if these are repeatedly inserted and removed. Continous insertion and removal may wear away the protective layer, allowing the underlying copper to oxidize, leading to loss of electrical conductivity. The finishes tend to be extremely flat, and are easily used with fine pitch devices. These surface finishes are already used in high-volume production today, and investigation into their optimization and further development continues.
The possible failure mechanisms described here should not serve to drive industrial users away from considering any of these alternative surface finishes, but any potential or current user should be aware of the common failure modes so that they can take action if necessary. The best ways to minimize the occurrence of these defects are firstly to choose a vendor that employs good process control, and secondly, to implement incoming batch inspection, either at your own facility or via a third party independent laboratory. There is of course additional cost associated with this, but in the long run it can save money - and your reputation - by preventing defective materials from entering your production process.
These new Pb-free finishes also have an effect upon solder joint metallurgy. We are moving from a primarily Sn-Pb system to a system that is predominantly tin, with the addition of perhaps four or five other elements, e.g. nickel and gold from the substrate, copper and silver from the solder paste, and palladium, nickel and gold from the component finish. These changes inevitably influence the intermetallic species observed and will also affect the way failure analysis is conducted on the solder joints(3). But that's a topic for a future column.
Brian J. Toleno is Application Engineering Team Leader with the electronics group of Henkel in Irvine, California. He holds a BS in chemistry and a PhD in analytical chemistry, and managed the EMPF's failure analysis laboratory in Philadelphia prior to joining Henkel.
Brian chairs the IPC's underfill handbook committee (J-STD-030) and co-chairs the solder paste standards committee (J-STD-005). He is also program chair for IEMT 2005, and an active member of SMTA. His published works include a course on failure analysis for SMTA, two chapters for electronic engineering handbooks, and numerous papers for trade publications.
1 - Biunno, N., IPC Expo March 1999, pp.S18-5-1
2 - Houghton, B. F. D., IPC Expo March 1999, pp.S18-4-1 and Houghton, B. F. D., IPC Works October 1999, pp.S-04-3
3 - Toleno B. et al., IPC-JEDEC Pb-free Conference, San Jose, 2004