Abstract:

In order to improve electronics packaging design, it is important to evaluate the cooling performance and reliability of the electronics packaging structure of a product. To that end, it is necessary to predict the temperature, deformation, and stress distributions of the package under field conditions. In the case of a packaging structure comprising a flip-chip ball grid array package, a heat spreader, thermal grease, a cooling structure, solder joints, and a motherboard, an increase in the contact thermal resistance may occur, depending on the interface contact condition between the cooling structure and the heat spreader due to the thermal deformation of the package. Contact thermal resistance problems involve the interactive relationship of the thermal and stress distributions. A coupled thermal-stress analysis, with consideration of the time-space variation of contact thermal resistance, was conducted to duplicate the behaviour of temperature, deformation, and stress distributions of a flip-chip ball grid array package under field conditions. It was found that: 1) the average contact thermal resistance across the interface between the heat-spreader and the plate fin, which was predicted by the coupled thermal-stress analysis, increased compared to the average contact thermal resistance in the case of uniform contact pressure, and 2) the contact thermal resistance will vary depending on the deformation mode, such as convex upward and downward, due to heat dissipation under field conditions. In addition, a reliability prediction method for thermal fatigue failure of solder bumps based on coupled thermal-stress analysis and statistical and probabilistic methods was proposed in order to select a suitable packaging solution at an early stage of design. It was found that the sensitivity of uncertain variables and the thermal fatigue life distribution of solder joints could change significantly depending on a combination of factors concerning the failure sites of solder bumps and the boundary conditions of the motherboard.

Index Terms—Anomaly detection, Bayesian linear models, Bayesian posterior class probabilities, kernel density estimation, one-class classifier, support vector machine.

Complete article is available to CALCE Consortium Members.

© IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.