Here is the abstract you requested from the IMAPS_2010 technical program page. This is the original abstract submitted by the author. Any changes to the technical content of the final manuscript published by IMAPS or the presentation that is given during the event is done by the author, not IMAPS.
|High Yield, Near Void-Free Assembly Process of a Flip Chip in Package Using No-Flow Underfill|
|Keywords: flip chip, yield, reliability|
|The advanced assembly process for a flip chip in package (FCIP) using no-flow underfill material presents challenges with high I/O density (over 3000 I/O) and fine-pitch (down to 150 µm) interconnect applications because it has narrowed the feasible assembly process of high electrical interconnect yield. In spite of such challenges, a high (>99.99%), reliable (sample size: 30) yield and nearly void-free (≈ 7.0%) assembly process has been achieved using commercial no-flow underfill material with a high I/O, fine pitch FCIP in our past research. The existing 7% void area could cause early failures such solders fatigue cracking or solder bridge in thermal reliability. Therefore, this study reviewed a classical bubble nucleation theory to predict the conditions of voids nucleation in assembly process. On the models prediction, systematic experiments were designed to eliminate underfill voiding using parametric studies. First, a void formation study investigated the effect of reflow parameter on underfill voiding and found process conditions of void-free assembly with robust interconnections. Second, a void formation characterization validated the determined reflow conditions to achieve a high yield and void-free assembly process, and the stability of assembly process using a large scale of assemblies respectively. Thus, this paper presents systematic studies into void formation study and void formation characterization through the use of structured experimentation which was designed to achieve a high, stable yield and void-free assembly process on the classical bubble nucleation. Indeed, the theoretical models were a good agreement in experimental results with a small discrepancy.|
|Daniel F. Baldwin, Ph.D.,