A two-day workshop, sponsored by the Metallurgy Division of MSEL, was held on April 17-18, 2008, to assess the state of the art of measurement methods of residual stress in lead-free electrodeposits used for surface finishes of electronic components. Thirty researchers from industry, government labs and universities attended.
In the 1960's the addition of 3% lead (Pb) to tin (Sn) electrodeposits was employed to eliminate "tin whisker" growth that caused shorts in electronic devises. Tin whiskers are typically from one to several micrometers in diameter and range from 10 micrometers to 2 millimeters in length. The increasing use of environmentally-friendly lead-free solder finishes has caused the problem to resurface. Researchers at NIST and elsewhere have shown that compressive stress in the electrodeposits combined with a columnar grain structure causes whisker growth. The accurate measurement of the state of stress and its time evolution in tin deposits with thicknesses between 1 micrometer and 10 micrometers has been recognized as necessary by industry, and the International Electronics Manufacturing Initiative (iNEMI) has asked NIST to develop these measurement methods for tin coatings. These measurements provide an essential component of fundamental research to understand and mitigate whisker growth. A robust method could also serve as a quality control method in industrial processes.
The benefits and difficulties with both X-ray diffraction (sine squared psi method) and wafer/beam curvature methods of stress measurement were discussed at the workshop. Although it is easy to obtain stress values from these techniques, understanding the source of measurement errors is undeveloped, and errors are often a significant fraction of the measured quantity. Sources of error include the variable preferred crystallographic orientation of the tin that arises during plating, and "creep", a time-dependent deformation prevalent in tin. The workshop addressed the following topics: precision and accuracy of measurements; interpretation of measurements, e.g. spatial resolution, macro- and micro- stress; orientation (texture) effects; grain shape effects, stress gradients in the deposit; and tin-copper intermetallic growth induced stress. Possibilities for local (micrometer scale) measurement of stress with synchrotron x-rays were also debated. Industry representatives encouraged NIST to publish a recommended practice guide on the subject. The discussions at the workshop have stimulated several research teams to collaborate and try new approaches to this measurement issue.