By Tim Ellis
In the previous edition of Emphasis I introduced Tin (Sn) whiskers as a reliability issue in electronics assembled using Pb - Free solders, as all are essentially based on Sn metallurgy. Along with showing a micrograph of Sn whiskers, several weapons systems were identified which have suffered failures from Sn whiskers in the recent past, e.g. Patriot, F - 15 Radar, Phoenix Missile (1, 2, 3, 4, 5, 6). Since electronic failure of a deployed system is not an acceptable option, what then is known about predicting failure due to Sn whisker growth?
A study by Dunn in the mid - 80's investigated the length of whiskers formed during room temperature storage, to Sn plated stressed Brass and Steel C - Ring samples (7). Three different Sn plate types were used: Type I, a "Normal" commercial process; Type II, an "Abnormal" high current density method; and Type III, where "Contamination" by the addition of organics was tested. Analyzing the data obtained from Sn plating on a Brass substrate, we find the activation energy, Q (the thermodynamic force to be overcome for growth to occur), varies with whisker length, Figure 1. Therefore, we can directly conclude that the growth of Sn whiskers is not a simple diffusion problem. Closer inspection of the data shows some interesting relationships, Table1.
What is observed is that the plating conditions have a large effect on Q, which is again the activation energy for whisker growth. Additionally, the value of Q found in these experiments is in reasonable agreement with the value given for the self-diffusion of Sn, 25.3 Kcal/Mole (8). Also, since the activation energy is lowest for Normal plating, one would expect that this coating should be the least susceptible to whisker growth.
Inspection of the data in Table 1 also shows that plating conditions may have a greater effect on whisker formation than imposed mechanical loads. This is not surprising as electro and/or electroless deposition are atomic transport processes that can develop high internal stress due to dislocations, contaminates, crystal structure selection and epitaxial issues. Since mechanical loading is not atomic in nature but essentially a macro-continuum, loading stress that develops is usually much lower in magnitude.
Work on-going here at ACI is to mitigate the effect of Sn whisker in high reliability systems required by the military. "Even though at the present time no definitive failure model exist for Sn whiskers maintenance and rebuild schedules can be developed from published experimental data." With the Kolmogorov - Johnson - Mehl - Avrami construct (9) ACI has developed a time-based failure criteria for Sn whisker growth in which the Growth coefficients for a specific alloy system is being determined empirically for presently available data, Equation 1:
Equation 1:
f = 1 - exp (Ktn) = fraction of material transformed to a whisker
K = constant proportional to self - diffusion and mechanical strain
n = Growth Exponent
t = Time
Exercising this simple model we can mimic the experimental data.
One result obtained using MatLab is shown in Figure 2, where we compare the different activation energies found in analyzing the data presented by Dunn.
This analysis is useful as the sensitivity of whisker growth to diffusion temperature and the value of the growth exponent, all well known, can be used to predict whisker length and therefore time to failure due to shorting. "Simply the spacing of the interconnects using Sn based material in microns is locate on the vertical axis of Figure 2. Moving across the diagram when the interconnection pitch distance value intercepts the line we drop down to the X - axis to find the time necessary to grow a whisker of that length. If the interconnection pitch distance is always above the line one would not expect electrical shorts due to whiskers." Therefore, some predictions can be made as to the expected length of service of Pb - free components can be made.
In the final installment of the EMPF's expose' on Sn whiskers the implications of this work on the reliability testing methods and assembly processes will be discussed. Along with a roadmap of the EMPF's future work in the Sn whisker arena.
References:
1.Military Airplane: G. Davy, "Relay Failure Caused by Tin Whiskers", Northrop Grumman Electronic Systems Technical Article, October 2002
2.Patriot Missile: Anoplate WWW Site: Suspected tin whisker related problems (Fall 2000)
3.Phoenix Air to Air Missile: L. Corbid, "Constraints on the Use of Tin Plate in Miniature Electronic Circuits", Proceedings 3rd International SAMPE Electronics Conference, pp. 773-779, June 20-22, 1989.
4.F-15 Radar: B. Nordwall, "Air Force Links Radar Problems to Growth of Tin Whiskers", Aviation Week and Space Technology, June, 20, 1986, pp. 65-70
5.U.S. Missile Program: J. Richardson, and B. Lasley, "Tin Whisker Initiated Vacuum Metal Arcing in Spacecraft Electronics," Proceedings 1992 Government Microcircuit Applications Conference, Vol. XVIII, pp. 119 - 122, November 10 - 12, 1992.
6.U.S. Missile Program: K Heutel and R. Vetter, "Problem Notification: Tin Whisker growth in electronic assemblies", Feb. 19, 1988, memorandum
7.A Laboratory Study of Tin Whisker Growth, B.D. Dunn, European Space Agency, ESA STR - 223, Sept 1987
8.CRC Handbook, 71st Edition, Radiotracer Diffusion Data for Pure Metals, 12 - 113, David R. Lide ed.
9.The Science and Engineering of Materials, Donald Askeland, 3rd Ed. , PWS Publishing, Boston MA (1994)
