Abstract

Stacked Chip Scale Package (SCSP) becomes popular recently because it increases package density and offers flexible component design at low cost. However, package reliability issues need to be carefully addressed because of the complex package geometry and structure. As one of the common failure modes, delamination is often encountered due to pre-existing defect and external loadings, such as thermal stress caused by different CTE of materials, hydro-swelling stress and vapor pressure caused by moisture desorption during reflow soldering. In SCSP, delamination often occurs along the interface of two dissimilar materials, primarily due to the weak interface bonding. Therefore, proper material selection and optimization of package structure are critical to reduce the risk of delamination in the design stage. Material selection is mainly based on the measurement of the interfacial fracture toughness or some quick-turn test to determine interface strength. Usually, the interfaces between die attach/die, die attach/substrate, mold/substrate and mold/die are evaluated. On the other hand, structural optimization is mainly carried out through numerical modeling to investigate the delamination sensitivity to stress, flaw size and position. Design parameters, such as die overhang, die thickness and bonding layer thickness are considered to have impact on the local stress distribution, which can lead to different delamination performance. Fracture mechanics based modeling, which can deal with pre-exist cracks, is adopted for this study. A typical multi-die stack chip scale package (CSP) is modeled by finite element to study the relationship between design parameters (die overhang, die thickness and bonding layer thickness) and stress intensity factor (SIF) for a delamination along die to mold compound interface. The effect of bonding layer (paste or adhesive film) CTE is also included in the present research as practically it can be changed by choosing different DA (die attach) materials. Optimized design parameters and considerations in package design can be proposed based on the modeling results.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call