Date of Graduation

12-2018

Document Type

Thesis

Degree Name

Master of Science in Mechanical Engineering (MSME)

Degree Level

Graduate

Department

Mechanical Engineering

Advisor/Mentor

David R. Huitink

Committee Member

Xiangbo Meng

Second Committee Member

Simon Ang

Third Committee Member

Ashok Saxena

Keywords

Electronic Packaging, Fatigue, Interconnects, Reliability, Solder, Thermo-mechanical stress

Abstract

The underlying physics of failure are critical in assessing the long term reliability of power packages in their intended field applications, yet traditional reliability determination methods are largely inadequate when considering thermomechanical failures. With current reliability determination methods, long test durations, high costs, and a conglomerate of concurrent reliability degrading threat factors make effective understanding of device reliability difficult and expensive. In this work, an alternative reliability testing apparatus and associated protocol was developed to address these concerns; targeting rapid testing times with minimal cost while preserving fatigue life prediction accuracy. Two test stands were fabricated to evaluate device reliability at high frequency (60 cycles/minute) with the first being a single-directional unit capable of exerting large forces (up to 20 N) on solder interconnects in one direction. The second test stand was developed to allow for bi-directional application of stress and the integration of an oven to enable testing at elevated steady-state temperatures. Given the high frequency of testing, elevated temperatures are used to emulate the effects of creep on solder fatigue lifetime. Utilizing the mechanical force of springs to apply shear loads to solder interconnects within the devices, the reliability of a given device to withstand repeated cycling was studied using resistance monitoring techniques to detect the number of cycles-to-failure (CTF). Resistance monitoring was performed using specially designed and fabricated, device analogous test vehicles assembled with the ability to monitor circuit resistance in situ. When a resistance rise of 30 % was recorded, the device was said to have failed. A mathematical method for quantifying the plastic work density (amount of damage) sustained by the solder interconnects prior to failure was developed relying on the relationship between Hooke’s Law for springs and damage deflection to accurately assess the mechanical strength of tested devices.

Download

Share

COinS