BY GLENN A. RINNE, Unitive Electronics, Inc., an Amkor Company
Tin is a common component of solders because of its desirable properties such as low melting point, high diffusivity, low surface tension in the liquid phase, and others. It also possesses undesirable habits that are a concern for flip chip and wafer-level chip-scale packages (WLCSP), such as tin whisker growth, electromigration, and tin pest. The first two of these have been well discussed, but tin pest is a little-known and potentially regrettable habit of pure-tin and tin-rich materials.
Tin pest is technically described as the allotropic phase transformation of the β-phase of tin (Sn) to the α-phase. This occurs when the temperature decreases below 13.2°C, although the transformation rate is extremely slow at this temperature. The density difference between the two phases creates significant stress in the lattice and will destroy the mechanical integrity of the solder joint because α-Sn is friable (readily crumbled). This gradual transformation seems to require an incubation period to nucleate a grain of α-Sn on β-phase Sn. Alternatively, the α-grain can be nucleated by a process called inoculation – exposure of β-Sn to a small seed grain of α-Sn. Once nucleated, the α-Sn grain grows radially across the surface and into the bulk where its lower density causes a circular swelling on the surface of the Sn. Nucleation in the bulk has not been observed because the driving force cannot overcome the lattice strain caused by the lower density of the α-phase.
The transformation is known to be sensitive to temperature although various sources have reported different critical temperatures ranging from -20 to -50° C.1 Although there is some historical evidence of this phenomenon, it has yet to be conclusively observed in lead-free solder bumps. A thermodynamic analysis of this phase transformation may provide an explanation of the paucity of occurrence to date.
Tin, and its two allotropic forms (α and β), was known to ancients. Often cited as contemporary evidence of mankind’s encounter with this phase transformation is the story of Napoleon’s buttons – the uniform buttons on the Grand Armie d’ Napoleon – which are said to have crumbled in the harsh winter outside Moscow in 1812, contributing to Napoleon’s defeat. However, a review of the data in Minard’s chart1 (Figure 1) describing the campaign shows that the temperature probably wasn’t low enough to cause tin pest until the retreat was well underway. Thus, this story is more interesting to engineers for the data presented in Minard’s chart than the effect of tin pest.
Figure 1. Minard’s Chart
Other common citations include the destruction of tin organ pipes in Northern Europe and the failure of the kerosene cans in caches along Robert Scott’s route in his failed expedition to the South Pole in 1912. However, conclusive documentary evidence is lacking in both cases, so the evidence is anecdotal at best.
In 1939, the U.S. Army investigated materials that could create a seal between artillery ordnance and the gun barrel. Tin was considered an ideal choice because it is was soft enough not to scratch the gun barrel yet could deform to create a tight seal.3 Tin pest was shown to be a major concern for this application, and an investigation demonstrated that alloying the Sn with Bi eliminated the tin pest concern, but increased the hardness and decreased the elasticity to such an extent that the desirable seal properties were lost.
Occurrence in Electronics
Although the transformation of the tin-rich phase in lead-free solders from β-Sn to α-Sn has been reported,3,4 there is little evidence of this in solder joints. Even lead-free solder bumps stored at low temperatures fail to exhibit the phenomenon. One possible explanation for this can be demonstrated using thermodynamics and rate equations for the transformation.
Figure 2. Thermodynamic kinetic model.
The α-phase is the energetically favorable phase at temperatures below 13.2°C. At this transition temperature both phases can coexist. As the temperature decreases, the free energy difference, ∆G, intensifies, increasing the driving force for the transformation. It would be reasonable to expect the rate of transformation to increase as the temperature decreases. At the same time, however, lower temperatures will decrease the kinetics by decreasing the thermal energy of the atoms in the lattice, reducing the probability that a given atom will cross the inter-phase boundary. The net of these two competing effects can be estimated using a hybrid thermodynamic-kinetic model5 (Figure 2). The interface between two regions of different phases will move as the transformation progresses and this model predicts the rate at which this occurs.
Figure 3. Model for small values of undercooling.
By applying a set of assumptions,6 a model for small values of undercooling can be applied (Figure 3). The final term in this equation is the driving force which is solely a function of the amount of undercooling. When the assumptions are applied, we can estimate the velocity of interface movement as a function of ambient temperature. The result is shown in Figure 4. As this is a fairly simple model, the assumptions are merely approximations. Further research is needed to refine them and validate the model.
Figure 4. Interface velocity as a function of temperature.
The peak rate occurs at approximately -2.5°C and decreases rapidly as the temperature decreases. Thus, although exposure to very low temperature may be necessary for nucleation of the α-Sn phase, kinetics are not favorable at these temperatures. In other words, the effect is not accelerated by exposure to lower temperatures so accelerated testing for tin pest should first expose the product to low temperatures (e.g., -40°C) for a period to nucleate the α-phase and then expose the product to moderately low temperature (e.g., -2.5°C) for a prolonged period to maximize grain growth of the α-phase.
Tin pest is a potentially insidious problem for lead-free solders. The incubation effect means that the α-phase cannot be dependably nucleated with a single excursion below the critical temperature, or, for that matter, any reasonable number of excursions. Unlike most other failure mechanisms, tin pest cannot be accelerated by either elevated or very low temperatures. Therefore, it is very difficult to accumulate a statistically significant dataset for estimating the reliability hazard. Perhaps we will just have to wait for sufficient field experience. Napoleon acquired his supporting data in 1812 and Scott acquired his in 1912. When might we expect to acquire ours?
- Hanson, A.C., Inman, G.O., Limitations of Tin as a Packing Material: Allotropic Transformation, Industrial and Engineering Chemistry, vol. 31, no. 6, pp 662-663, June 1939.
- Minard, C. J. Carte figurative des pertes successives en hommes de l’armée française dans la campagne de Russie, 1812-1813, 20 November 1869. ENPC: Fol 10975, 10974/C612. From Friendly, M., The Graphic Works of Charles Joseph Minard, 2006, http://www.math.yorku.ca/SCS/Gallery/minbib.html.
- Kariya, Y., Williams, N., Gagg, C., Plumbridge, W., Tin Pest in Sn-0.5 wt.% Cu Lead-Free Solder, MRS Journal of Materials, June 2001.
- Kariya, Y., Gagg, C., Plumbridge, W., Tin Pest in Lead-free Solder, Soldering and Surface Mount Technology, v. 13, no. 1, pp. 39-40, 2000.
- Porter, D.A., Easterling, K.E., Phase Transformation in Metals and Alloys, Van Nostrand Reinhold, New York, 1981.
- Warnes, W.H., http://oregonstate.edu/instruct/me482/Homework/W06/ME482Hmwk4.html.
GLENN RINNE, VP R&D, may be contacted at Unitive Electronics, Inc. 140 Southcenter Ct. Suite 600, Morrisville, NC, 27560; 919/459-1204; E-mail: firstname.lastname@example.org.