Finite Element Modeling (FEM) has been widely used for analysis of tin-lead and lead-free solder reliability and prediction of their lifetimes. Many past research efforts have already modeled tin-lead solder reliability and given a relatively accurate prediction of lifetime reliability of tin-lead solder. Lead-free solder modeling and reliability prediction are still a great challenge to researchers and manufacturers.

1.) Lead-free solder failure modes will be different from tin-lead solder.
With tin-lead, grain coarsening is an indication of environmental condition; with lead-free solder, this may not be the case. Higher reflow temperature and different physical properties of the alloy may change the failure mechanism of the joints and surrounding areas. Investigations will need to include an expanded area when evaluating cross-sections.

2.) It is known that Ag makes the solder materials more brittle (higher brittleness transition temperature).
Even when creep will be lower at the higher temperature for lead-free solder materials, the lower temperature could be the crack initiator. The cracks begin to grow even in the first cycles, but only in the case where temperature was dropped to -55°C.

3.) New intermetallic regions are formed, which may result in early brittle fractures.
A good example is a SnPb solder joint on a NiAu-finished PCB. Samples failed at one-third of the intrinsic fatigue life of the joint itself. Can we expect similar problems with lead-free solder materials?

4.) Data will need to be gathered for many different types of lead-free solder materials in various compositions.
SnPb will not be replaced by a single lead-free solder material. Both material data (E-modulus, CTE, creep behavior) and correlation models (e.g., relation between creep strain and lifetime) must be measured for all new materials.

5.) How are the acceleration factors for lead-free solder materials?
W. Engelmaier determined that lead-free solders have creep rates up to 100 times slower than creep rates of standard SnPb solders. 1 The implication is that meaningful reliability tests cannot be significantly accelerated; while the use of lead-free solder for consumer goods like cell phones is acceptable, it clearly cannot as yet be recommended for high-reliability applications.

Experimental study results have generated no general conclusion about the trend in lifetime reliability from SnPb to SnAgCu. The trend is very dependent on the package type as well as the applied loading conditions. This means the lead-free reliability and lifetime prediction issues are very case-dependent; we cannot make universal or general conclusions.

Examples of research efforts
IBM (Farooq, et al) 2 developed a Finite Element Model (FEM) where the mesh consisted of a 3D slice of the module and card from the center of the array to the corner, encompassing a diagonal row of interconnections. The simulated structure was then subjected to a single 0-100°C thermal cycle. The plastic strain amplitude (Coffin-Manson) approach to fatigue was used.

Agilent, IBM, and Alcatel (Lau, et al) 3 constructed Finite Element Models for a 256/388 pin plastic ball grid array (PBGA) and a 1657 pin ceramic column grid array (CCGA). They claim the lead-free solder obeys the Garofalo-Arrhenius creep constitutive law.

The models include nonlinear material properties and a 3D strip that captures the construction along a diagonal path from the geometric center of the package to a corner. CALCE, University of Maryland (Blattau, et al) 4 used elastic-plastic finite element simulations to study the lead-free solder and flex cracking failure in MLCC (multi-layer ceramic capacitors). The solder attaching an MLCC is a critical path along which the printed wiring board (PWB) connects to the capacitor; therefore, the solder properties play an important role in the durability of ceramic chip capacitors.

The EMPF is using FEM to analyze the effect of the high Young’s modulus of lead-free solder on thermal stress/strain and reliability. The model includes substrate, copper pad, and solder ball. The model simulates thermal cycling testing from 0°C to 100°C and thermal shock testing from -55°C to 125°C. For the concept simulation, only one solder ball model is established. The transient thermal analysis is used with a 4000 1/s capture rate and the temperature file is input into MES (mechanical event simulation) to obtain stress/strain results. The Coffin-Manson method is used to estimate the lifetime
of the simulated solder joint.

There are still many questions and some conflicting data regarding lead-free reliability. Due to the highly case-dependent nature of the research, this is likely to continue.

References
1. “Lead Free Solder Joint Reliability Estimation by Finite Element Modeling Advantages, Challenges and Limitations,”
B. Vandevelde, et al, IMEC, Leuven, Belgium.
2. “Thermo-Mechanical Fatigue Reliability of Pb-Free Ceramic Ball Grid Arrays: Experimental Data and Lifetime Prediction Modeling,” Farooq, et al, IBM Microelectronics, Inter Connect Packaging Development, Hopewell Jct., New York.
3. “HDPUG’s Design for Lead-Free Solder Joint Reliability
of High-Density Packages,” Lau, et al, IPC SMEMA Council APEX 2003, www.GoAPEX.org
4. “Lead Free Solder and Flex Cracking Failures in Ceramic Capacitors,” N. Blattau, et al, CALCE EPSC, University of Maryland, College Park, MD

For more information concering Lead Free processes and surrounding issues, please stop by ACI's new Lead Free Manufacturing Page to download articles contributed to ACI by some of the industry's most knowledgable individuals and organizations, as well as material generated by ACI, and documents on the legislation surrounding the Lead Free issue.