Life Prediction Model of Electronics Subjected to Thermo-Mechanical Environments

Abstract

Microelectronic assemblies used in automotive, military and defense applications may be subjected to extreme high and extreme low temperature in addition to temperature cycling with intermittent prolonged period of storage. Field deployed electronics, unlike controlled laboratory testing, may experience a variety of sequential thermo-mechanical stresses during its lifetime. The field environment imposed on the electronic system may be influenced by ambient conditions and usage profile. Variation in the ambient temperature conditions, weather variations, usage location in addition to scheduled system usage or downtime may result in wide variance in the resulting field profile. Thermally induced stresses can be produced in fabrication stage also. In an electronic packaging assembly, one of the crucial parts affecting reliability the most is the solder joint, which actually bears most of the stress gradients formed inside packages due to CTE mismatch. Safety critical nature of the electronics systems necessitates that level-of-damage and reliability should be assessed prior to any future redeployment. The prognostic technique proposed in this thesis is based on Physics of Failure based damage indicators of second level interconnects. Both the test vehicle and thermal conditions have been thoughtfully selected to ensure relevancy to current packaging technologies and critical usage profiles. The impact of iso-thermal aging and effect of mean temperature of thermal cycling environment were investigated and the information derived from these investigations were used to develop the damage mapping and life prediction model. The damage mapping relationship deals with finding a Time-Temperature combination required to reach a particular damage state and the predictive model shows how microstructural growth rate is related to fatigue life. The evolution of two different microstructural damage parameters were used in the process of developing the life prediction model. Apart from microstructure based parameter, a new prognostic tool based on shear strain was presented, where shear strain evolution with number of thermal cycles undergone was investigated with a view to finally correlate with life degradation. A full-field optical technique like Digital Image Correlation was used to investigate the evolution of shear strain with number of thermal cycles, which is a completely new approach to assess the effect of thermal cycling duration on shear strain in real time. A quarter symmetry model of package was built in ANSYS and simulated using Anand’s Visco-plastic Constitutive Model to check the validity of experimental finding and at the same time to calculate plastic work dissipated under a particular thermal cycling environment. Finally popular life models like Coffin-Mansion and Darveaux’s model were implemented with experimental data. A new model of Remaining Useful Life (RUL) prognostication was proposed based on microstructural damage indicators. Also a new approach to fatigue life calculation was presented by correlating microstructural parameter growth and plastic work dissipated.