Creep corrosion and its impact on reliability
BY JINGSONG XIE AND MICHAEL PECHT
|
As a leadframe finish material, palladium has many advantages over tin-lead (SnPb) solder; it is more environmentally friendly and the package assembly process with palladium-plated leadframes is more efficient. However, palladium also provides a surface over which corrosion products can migrate and cause leakage current failure modes, especially in fine-pitch devices. The time-dependent migration of corrosion products is known as creep corrosion. Creep corrosion behavior is dependent on the corrosion product, surface material, and environmental and operating conditions.
Palladium's Evolution and Usage
Palladium-plated leadframe technology began in the 1980s, primarily to reduce manufacturing steps and cycle-time in package assembly; a single plating process can be used to ensure good die attachment and wirebonding, reduce lead corrosion, and enhance solderability to a PCB. The initial applications were for low pin-count, surface-mount packages, as well as some selected through-hole components. Today, palladium plated leadframes are being incorporated in many fine-pitch, high pin-count surface mount packages, such as thin quad flat pack packages (TQFPs) with more than 100 pin counts and a lead pitch of 0.4 mm.1-5
As an ongoing effort of many electronics manufacturers, removal of lead (Pb) from electronic components has become a new consideration for the adaptation of palladium lead finishes, especially as the cost of palladium has increased. In particular, the palladium leadframe plating process has been found to reduce workplace health and safety concerns and improve waste disposal.6 In Europe and Japan, changing to lead-free products is also considered a strong product differentiator and a potential future legislative mandate.
There are many different ways in which palladium is currently used as a leadframe plating material. Nickel-palladium (NiPd) plating, such as palladium over nickel over copper base metal, is the most fundamental structure. Usually, a top layer of palladium strike with thickness ranging from 0.025 to 0.127 µm protects the underlying nickel from corrosion. During the solder reflow process of surface mount components onto a PCB, the palladium dissolves into the solder, and solder joints are intermetallically bonded with the underlying nickel plate.1
Since the initial concept of the palladium lead finish was introduced, various qualification studies have been performed to assess palladium-plated parts and testing has been conducted to evaluate part shelf-life and impact of storage on solderability.6-14 These studies have generally concluded that a palladium lead finish can provide at least equivalent corrosion resistance capability, solderability and shelf life compared with the standard tin-lead. The biggest stated limitation has been that palladium coating is more prone to surface damage, and palladium-plated parts may not always pass standard “dip and look” solderability tests. Unfortunately, there have been very few studies of creep corrosion.
Creep Corrosion
Creep corrosion is a physical process during which solid corrosion products migrate over a surface. Usually the corrosion process initiates from some surface damage, such as a crack or scratch in the lead finish, which exposes the underlying copper leadframe.15 This can arise before, during or even after mounting a component onto a PCB. Often, the defect occurs at the lead bend. Dependent on the environmental conditions, corrosion products may be continuously generated from the defect site(s) and diffuse over the lead surface.
The surface diffusion process is dependent upon the chemical species of the corrosion products. A copper leadframe is chemically active and will oxidize if exposed to the air without protection, but this oxide species is not mobile. On the other hand, copper chloride has significantly higher surface mobility than copper oxides, and can significantly accelerate the creep corrosion of copper corrosion products. Chloride and sulfur species can continuously regenerate and are most conducive to creep corrosion. Mixed flowing gas (MFG) tests are often conducted to assess creep corrosion based on various combinations and concentrations of corrosive gases.
Figure 1. Schematic diagram of corrosion products migrating over palladium-plated leads and across the plastic molding compound, thus electrically bridging the leads. |
A “surface diffusion coefficient” is used to quantify the mobility of corrosion products over a surface under given environmental conditions. A high surface diffusion coefficient represents a material with low resistance to creep corrosion. Both palladium and gold have high surface diffusion coefficients. However, while the mechanisms of creep corrosion over palladium and gold surfaces are similar, palladium surfaces tend to have higher surface resistance to creep corrosion than gold surfaces because palladium develops a few atomic layers of surface oxide when exposed to an ambient environment.16 In fact, gold has the lowest resistance to creep corrosion among metals. Experimental observations have shown that copper sulfide creeps over gold at an average rate of 0.005 to 0.013 mm/hr.17-20 Creep corrosion can also occur on plastic surfaces, although no one has yet benchmarked this effect with respect to metals.
Creep corrosion is driven by concentration gradients of chemical species of the corrosion product. The chemical species move from the areas with higher concentration to those with lower concentration. Increasing the relative humidity and temperature generally leads to increased mobility of diffusing substance and the availability of moisture to ionize the diffusing solid.
Reliability Concerns and Mechanisms of Current Leakage
In a plastic encapsulated leaded surface mount package, the failure mechanism called creep corrosion will only be a reliability concern if the corrosion product is electrically conductive and bridges across two electrical paths, such as leads. The failure mode is generally current leakage.
Figure 1 diagrams the process in which corrosion products migrate over the palladium-plated leads and across the plastic molding compound to bridge the leads. Here, the corrosion products are complicated mixtures, which can have some degree of electrical conductivity, especially if ion conduction-based corrosion is involved. In humid conditions, the formation of electrolytes can ionize corrosion products and increase the mobility of ions if the corrosion products involve water-based reactions. As a result, just like conduction in an electrolyte, corrosion products, which bridge adjacent leads, can cause leakage current between the I/Os.
Figure 2. Typical palladium-plating structures for leadframes. |
It turns out that the plastic encapsulation also plays a role in the current leakage problem, because the corrosion product must bridge across the mold compound to create a path between two leads. This is possible because the plastic can also enable creep corrosion and because there may be a voltage bias between the leads, which can accelerate the migration of corrosion products. A somewhat counter- intuitive observation is that a matted plastic surface appears to be more prone to creep corrosion than a smooth surface. In general, the surface condition of the plastic is determined by requirements on part marking. For example, a stamping process is usually used for printing on a smooth plastic surface, while laser printing generally requires a matted surface. The CALCE Electronic Products and Systems Center at the University of Maryland is currently conducting experiments to understand this phenomenon.
Palladium and Gold
Gold and palladium are two of the most widely used materials for electronics plating applications. Palladium has a higher protection ability than gold with the same coating thickness. In fact, the surface porosity of palladium plating is approximately one-fifth that of gold plating.15 Nevertheless, the enthusiasm for palladium as a lead finish has dampened as the price of palladium has dramatically increased. Palladium was about a quarter of the price of gold in 1980, less than one-third the price of gold in 1990, but by the end of 2000 it was more than two times the price of gold. As a result, an alternative to palladium-plating as a lead-free solution is to replace the palladium with gold or a combination of gold and palladium. For example, Texas Instruments has recently begun to follow many Japanese companies in using nickel-palladium-gold (NiPdAu) plating for selected devices.21
Acknowledgements
Research into semiconductor and MEMs packaging, and the use of plating materials and creep corrosion is still ongoing at the CALCE Electronic Products and Systems Center of the University of Maryland. The research is supported by more than 100 electronic product and systems companies, including telecommunications, computer, avionics, automotive and military electronics manufacturers.
References
- M. Pecht, L. Nguyen, and E. Hakim, Plastic Encapsulated Microelectronics, John Wiley Publishing Co., New York, NY, 1995.
- M. Pecht, R. Radojcic and G. Rao, Guidebook for Managing Silicon Chip Reliability, CRC Press, Boca Raton, FL, 1999.
- M. Pecht, et al., Electronic Packaging Materials and their Properties, CRC Press, Boca Raton, FL, 1999.
- D.W. Romm, “Palladium Lead Finish User’s Manual,” Texas Instruments, 1998.
- J.C. Yang et al., “Palladium Pre-plated Copper Lead Frame for DRAM LOC Packages,” 1999 Proceedings of the 49th Electronic Components and Technology Conference, San Diego, CA, June 1999, pp. 842-847.
- D.C. Abbott, “Nickel/Palladium Finish for Leadframes,” IEEE Transactions on Components and Packaging Technology, 22 (1), March 1999, pp. 99-103.
- E.E. Benedetto, “Evaluation of Palladium Lead Plating,” SMT, 13 (2), February 1999, pp. 90-96.
- E.E. Benedetto, “Solderability and Board-Level Reliability of Palladium-Plated, Fine-Pitch Components,” Proceedings of Surface Mount International Conference and Exhibition, San Jose, CA, August 1998, pp. 624-631.
- D.W. Romm and D.C. Abbott, “Steam-Age Evaluation of Nickel/Palladium Lead Finish for Integrated Circuits,” Texas Instruments, September 1998.
- D.C. Abbott, R.A. Frechette, G. Haynes and D. Romm, “Shelf-life Evaluation of Nickel/Palladium Lead Finish for Integrated Circuits,” Texas Instruments, April 1998.
- D.W. Romm and D.C. Abbott, “Pb-Free Solder Joint Evaluation,” SMT, March 1998, pp. 84-88.
- D.W. Romm and W.R. Reynolds, “Surface Mount Solderability Test,” SMT, January 1994.
- D. Romm and N. McLellan, “Water-Soluble vs. No-Clean Pastes: The Factors that Affect Performance,” SMT, January 1993, pp. 35- 40.
- D.C. Abbott, et al., “Palladium as a Lead Finish for Surface Mount Integrated Circuit Packages,” IEEE Transactions on Components and Packaging Technology, Vol. 14, No. 3, September 1991, pp. 567-572.
- R. Martens and M. Pecht, “An Investigation of the Electrical Contact Resistance of Corroded Pore Sites on Gold Plated Surfaces,” IEEE Transactions on Advanced Packaging, Vol. 23, No. 3, Aug. 2000, pp. 561-567.
- I. Memis, “Enabling Grid Array Modules Through Advanced Printed Wiring Board Surface Finish,” MicroNews, Vol. 5, No. 4, www.chips.ibm.com/micronews, 1999.
- L.R. Conrad, M.J. Pike-Biequnski and R.L. Freed, “Creep Corrosion over Gold, Palladium and Tin-Lead Electroplate,” 15th Annual Connectors and Interconnection Technology Symposium Proceedings, Philadelphia, PA, November 1982, pp. 401-414.
- V. Tierney, “The Nature and Rate of Creep of Copper Sulfide Tarnish Film over Gold,” J. Electrochem. Society: Solid-State Science and Technology, June 1981, pp. 1321-1326.
- J. Xie, M. Sun, M. Pecht and D.F. Barbe, “Why Gold Flash Can Be Detrimental to Long-Term Reliability,” ASME Journal of Electronic Packaging 123.3.
- W. H. Abbott, “Corrosion of porous gold plating in field and laboratory environments,” Plating and Surface Finishing, Vol. 74, No. 11, 1987 pp. 72-75.
- Texas Instruments, Product/Process Change Notification: 20000926002: Notification of Additional Lead Finish (NiPdAu) DIP (8P, 14/16/20N, 24NT, 16/20NE) Pkgs. – MSLP, (available from https://mist.ext.ti.com/pcn) November 11, 2000.
null
JINGSONG XIE is a research engineer at Microsoft, Redmond, WA 98052, and MICHAEL PECHT is a professor and the director of the CALCE Electronic Products and Systems Center, University of Maryland, College Park, MD 20742; 301-405-5323; Fax: 301-314-9269; E-mail: [email protected].