MEMS Cantilever on High-Cycle Fatigue Testing of thin Metal Films

N. Jöhrmann, C. Stöckel,B. Wunderle

2022 23rd International Conference on Thermal, Mechanical and Multi-Physics Simulation and Experiments in Microelectronics and Microsystems (EuroSimE)(2023)

引用 0|浏览0
暂无评分
摘要
Aluminium is still one of the most important contact metallisation for power electronic chips like MOSFETs or IGBTs. With a large difference in thermal expansion coefficients (CTEs) between aluminium and silicon or silicon carbide, and the temperatures generated in hot-spots during high power transients, these layers are prone to failure due to thermo-mechanical fatigue. Usually lifetime assessment is done by subjecting dedicated test specimens to standardised stress tests as e.g. thermal cycling. This paper builds on previous work about a method for accelerated stress testing and lifetime modelling of thin aluminium films in the high-cycle fatigue regime by isothermal mechanical loads using Si MEMS cantilevers as sample carriers. Surface roughness of the fatigued films was previously measured as a failure parameter, both via atomic force microscopy (AFM) and scanning electron microscopy (SEM). This was motivated within a physics-of-failure based reliability paradigm by comparison with equivalent plastic strain obtained from finite element simulations, which show a relation between surface roughness and the cumulated plastic strain. By using a new design including two AlN piezos, it is now possible to realize a closed loop control without the need for an external shaker and an optical setup to measure the cantilever amplitude during the fatiguing. This also enables easier Insitu stress testing, e.g. inside a SEM or below an optical microscope, to gain further insight into the development of degradation of the thin aluminium over time. An accelerated stress test with 10 7 cycles using the new design is presented, and surface roughness obtained via optical microscopy is compared with finite element simulations.
更多
查看译文
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要