Reading time ( words)
This article presents results of an eight-month assessment of whisker growth on matte-tin-plated copper in the presence of a constant electrical current. Results show a reduction in whisker density due to the application of an electrical current, but a slight increase in whisker length.
By Yuki Fukuda, Ph.D.; Michael Osterman, Ph.D.; and Michael Pecht, Ph.D.
The electronics industry is in the process of eliminating lead from electrical and electronic products - driven by legislative requirements and market forces.1 A significant portion of electronic-part manufacturers have adopted pure-tin and tin-rich alloys as lead-free options due to their low cost and compatibility with existing solders. However, the adoption of high-tin-content finishes has created a reliability issue pertaining to the formation of conductive tin whiskers, which can bridge adjacent conductors and result in current leakage or electrical shorts.
There have only been two studies on the role of an electrical current on tin whisker growth, and these have both involved evaluations of less than 1,000 hours (less than two months). One experiment was conducted on matte-tin-plated brass using four different levels of current density (0.25 × 102 to 3.12 × 102 A/cm2) under three exposure conditions.2 The study concluded that current flow does not affect whisker growth. A separate study was conducted in which accelerated electromigration tests used a Blech structure of 5,000 Å (=0.5 µm) pure tin over 700 Å titanium. In this study, whisker growth was observed at the anode end, while tin-grain depletion increased at the cathode end with current stressing time and temperature.3 It was concluded that whiskers grow to release compressive stresses generated by tin-atom movements from cathode to anode due to electromigration. The growth rate of tin whiskers was reported as 3 Å and 7.7 Å/sec. at room temperature and 50ºC, respectively, in the presence of current density of 1.5 × 105 A/cm2. However, the applied current density, tin-plating thickness, and base material were not representative of any actual electronic component applications.
This study examines whisker growth under the application of an electrical current on matte-tin-plated copper, the most widely used material combination for electronic components. To assess whisker growth under practical electrical current stressing, annealed and non-annealed samples were subjected to 0.48 × 102 A/cm2 for eight months.
Copper metal coupons with dimensions of 12.7 × 31.7 × 0.15 mm were commercially electroplated on all sides with matte tin, with a measured average thickness of 5 µm (±0.4 µm). To simulate the trim-and-form process, plated samples were bent approximately 90° at both ends over a plastic fixture (Figure 1). A constant electrical current density of 0.48 × 102 A/cm2 was then applied to half of the samples; the other samples served as controls. This is typical of the current density of power electronics, such as power converters for wireless network access and microprocessor-powered applications. Three samples per test condition were then placed in a temperature humidity chamber at 50ºC/50% relative humidity (RH) for eight months. This exposure condition was chosen to accelerate the whisker formation, based on studies reporting the higher whisker propensity at these conditions.4
Figure 1. Bending method - this method applies to both sides of the tin-plated copper coupons.
Measurements of maximum observed whisker length, length distribution, and whisker density were taken initially and weekly for up to eight months of exposure to the electrical current and temperature-humidity (50ºC/50%RH) conditions. Fifteen surface-site characterizations (one site: 725 × 625 µm) were conducted. An environmental scanning electron microscope (ESEM) was used to accurately measure whisker density and calculate whisker length at the flat area, inner-curved area, and outer-curved area of each sample (Figure 2).
Figure 2. Applied voltage and observation areas.
Observations and Discussions
The surface morphology of as-received tin-finishes, consisting of both annealed and non-annealed samples, was observed under ESEM to ensure grain sizes were within accepted specifications for matte tin.5 Whiskers were observed on all samples, and at all areas. However, whiskers tended to be more densely populated and longer at the inner-curved area, compared to the flat surface areas. This is expected because applied mechanical bending causes compressive stresses at the inner-curved surfaces. All data sets of whisker-density data were best fitted by a lognormal distribution.
The axial lengths of the tin whiskers were measured as the distance between the tin surface and the tip of the whisker, per JESD-22A-121.6 For bent whiskers, the total axial length was estimated by adding all of the straight sub-divisions of a whisker. Curved whiskers were divided into piece-wise linear elements. Whisker growth was observed to initiate approximately three to five weeks after electroplating, based on weekly surface observations. The initiation of whisker growth at the bend area occurred about two weeks sooner than those at the flat surface area.
Contrary to a previous study,3 whisker growth was observed at both the anode and cathode ends of the tin-plated samples. No discernible voids or depletion of tin grains at the cathode end were observed under ESEM. This could be due to the much lower level of current density adopted in this experiment. The previous experiment focused on electromigration phenomenon with current densities of 1.5 × 105 A/cm2.
Figure 3. Whisker density of matte-tin sample with one standard deviation, with and without annealing and the application of electrical current after eight months.
Figure 3 shows the whisker density observed at eight months at the coupon flat area. The graph shows the mean whisker density with one standard deviation on both sides. Compared to the control samples (i.e., non-annealed and no current applied), the application of electrical current resulted in a lower whisker density. On the other hand, the presence of the electrical current increased the maximum whisker length in the samples slightly (Table 1). In this case, maximum whisker length is defined as the 99-percentile value of the lognormal distribution fitted to the experimental data. Therefore, although the electrical current reduces the whisker density, the maximum length of whiskers will increase.
Table 1. The 99-percentile value of maximum whisker length (μm) over an eight-month period on the flat surface of a test coupon.
1 Ganesan, S., and M. Pecht, “Lead-free Electronics,” John Wiley and Sons, Inc., 2006.
2 Hilty, R. D., N. E. Corman, and H. Herrman, “Electrostatic Fields and Current-flow Impact on Whisker Growth,” IEEE Transactions on Electronics Packaging Manufacturing, Vol. 28, No. 1, pp. 75-84, January 2005.
3 Liu, S. H., C. Chen, P. C. Liu, and T. Chou, “Tin Whisker Growth Driven by Electrical Currents,” Journal of Applied Physics, Vol. 95, No. 12, pp. 7742-7747, June 2004.
4 Woodrow, T., and E. A. Ledbury, “Evaluation of Conformal Coating as Tin Whisker Mitigation Strategy,” IPC/JEDEC 8th Int’l Conference on Lead-free Electronic Components and Assemblies, April 2005.
5 “Recommendations on Lead-free Finishes for Components Used in High-Reliability Products,” iNEMI, May 2005.
6 “Measuring Whisker Growth on Tin and Tin Alloy Surface Finishes,” JEDEC Standard 22A121, May 2005.
Yuki Fukuda, Ph.D., formerly of CALCE, reliability engineer, Philips Lumileds Lighting, may be contacted via e-mail: firstname.lastname@example.org. Michael Osterman, Ph.D., director, CALCE, may be contacted via e-mail: email@example.com. Michael Pecht, Ph.D., chair professor and director, CALCE, may be contacted via e-mail: firstname.lastname@example.org.