Investigation of bulk solder and intermetallic failures in lead-free BGA by joint level testing

Concerns about potential, in particular uncontrolled, failure account for a very large fraction of the efforts and costs associated with the packaging of high-end microelectronics products. In the case of high liability, products for the military, aerospace, and medical sectors the reasons for this are obvious. Still, great efforts are often invested in the optimization of the reliability of consumer products where a bad reputation can have long lasting effects on sales.;When it comes to SMT type assembly technologies a major factor is the robustness and reliability of the interconnects, most commonly the solder joints. Depending on product design, materials selection and service conditions the latter may fail within the solder itself, along or within the intermetallic structure bonding it to one of the contact pads, by delamination of a pad from the underlying laminate resin, or by 'cratering' through this resin.;Failure of solder joints may be divided into two types, those associated with defects or imperfections, and those caused by the overstressing or wear-out of good joints. These are dealt with in very different ways. The failure of a properly optimized solder joint due to thermal mismatch induced stresses should, for example, always occur wholly within the solder. Failures within an intermetallic bond or by cratering should not be accounted for in the prediction of service life. Instead, they should be prevented by design optimization and/or elimination of defects. Solder failure in thermal cycling, on the other hand, is in principle fully understandable and predictable, allowing for both optimization and assessment of the useful product life under given service conditions. Isothermal drop, vibration, bending, etc. often leads to intermetallic failure or pad cratering, even in the absence of defects, so here these phenomena must be accounted for in optimization and assessment.;If the root cause is not already known, the prevention of a particular type of defect may require a massive research effort. One goal of the present work was to help define and focus such an effort for a sporadic, but occasionally catastrophic, intermetallic problem. Otherwise, the optimization of design and materials selection often requires at least the qualitative assessment of robustness and reliability, and a quantitative understanding is obviously needed for the prediction of actual service life. Over the past several decades such an understanding has to a large, albeit incomplete, extent been established for conventional applications of SnPb solders. However, the forced transition of most electronics products to lead free soldering has provided for a major setback. In fact, the implicit reliance on assumptions and procedures developed for Sn-Pb solder, even if modified somewhat, continues to provide for serious errors and misperceptions for lead free. Addressing this in a comprehensive fashion is far beyond the scope of the present work, or indeed the capabilities of any single researcher. Still, another goal of this thesis is to contribute significantly to the understanding of accelerated tests and test results.