Hygrothermal failure is a long existing issue for plastic encapsulated IC packages. The moisture absorbed by plastic packages will vaporize and produce high internal pressure during the reflow heating process, in which a peak temperature typically ranges from 220°C to 260°C. As a result, the popcorning crack/interfacial delamination may occur under the combined effects of thermal stress, vapor pressure and interfacial strength degradation. These hygrothermal failures can damage the electronic interconnections and cause short term or long term reliability issues. The “bake-n-bag” and labelling method has been widely used in the industry to manage the aforementioned failures. However, the physics of failure and the detailed cracking mechanism have not yet been fully understood due to the...[ Read more ]

Hygrothermal failure is a long existing issue for plastic encapsulated IC packages. The moisture absorbed by plastic packages will vaporize and produce high internal pressure during the reflow heating process, in which a peak temperature typically ranges from 220°C to 260°C. As a result, the popcorning crack/interfacial delamination may occur under the combined effects of thermal stress, vapor pressure and interfacial strength degradation. These hygrothermal failures can damage the electronic interconnections and cause short term or long term reliability issues. The “bake-n-bag” and labelling method has been widely used in the industry to manage the aforementioned failures. However, the physics of failure and the detailed cracking mechanism have not yet been fully understood due to the inadequate experimental investigations and the premature moisture studies.

This thesis was devoted to implement a comprehensive study on hygrothermal failure in electronic packages including both of detailed experimental and computational works. A relatively new and popular type of electronic package, namely, quad flat no-lead (QFN) package, was adopted for demonstration. The primary and advanced studies on initial cracking analysis were conducted on dummy and industrial QFN packages, respectively. A number of experimental methods and computational techniques such as moisture sensitivity level tests, C-SAM inspection, cross-section analysis, high temperature button shear/pull tests, and finite element analysis were carried out to explain the initial cracking mechanism. It was observed that the failure always initiates at the molding compound/lead-frame interface around the junction of die attach fillet. Besides, the thermal stress is the dominant stress and the hygroswelling effect is insignificant.

Once the crack initiates in packages, moisture can transfuse from the molding compound into the initial crack opening zone (COZ) due to the nature of mass diffusion and produce vapor pressure at cracking surfaces. A 1-D moisture transfusion model was then proposed to study the rate of moisture transfusion and the related vapor pressure in COZ. Differing from the conventional models, the newly developed transfusion model started from an equation similar to the 1-D Newton’s Law of cooling for natural heat convection. The analytical solution of transfusion model showed that the vapor pressure in COZ is a function of saturated moisture concentration in molding compound, temperature, time, the “moisture convection” coefficient, and the crack opening.

Subsequently, the crack propagation study was implemented on finite element models with the vapor pressure acting on embedded cracking surfaces. The strain energy release rates were calculated by crack tip opening displacement techniques. From the comparison with MSL test results, it was observed that the strain energy release rates calculated from the current transfusion model are more reasonable than those calculated from the conventional model. Moreover, the ranges of critical strain energy release rate were evaluated by comparing the calculated strain energy release rates to the MSL tests. Thus, the crack propagation in QFN packages under targeted MSL tests became predictable eventually.