SUBJECT: Ph.D. Proposal Presentation
   
BY: Eva Baumert
   
TIME: Friday, September 14, 2012, 1:00 p.m.
   
PLACE: Love Building, 210
   
TITLE: Influence of Environment on the Fatigue Properties of Silicon and Nickel Films
   
COMMITTEE: Dr. Olivier Pierron, Chair (ME)
Dr. Stephen Antolovich (ME)
Dr. Oliver Brand (ECE)
Dr. Samuel Graham (ME)
Dr. David McDowell (ME)
 

SUMMARY

The fatigue behavior of two diff erent classes of structurally used materials in microelectromechanical systems (MEMS), namely silicon and nickel, is investigated. Previous research has shown that environment aff ects the behavior of both materials. Specifically, surface oxidation has been found to be key in the fatigue of both silicon and nickel. The goal of this study is to reveal further insight into the underlying mechanical behavior of both materials at the small scale. This understanding is necessary to improve the reliability of silicon and nickel MEMS devices in the future.
In cyclic fatigue testing, micro-resonators were subjected to fully-reversed fatigue testing at resonance (40 kHz for silicon, 8 kHz for nickel). Experiments were conducted in air at 30 °C, 50% relative humidity (RH) or 80 °C, 90% RH and testing was carried out at a broad range of applied stresses. Tracking of the resonance frequency was used as a direct measure of the fatigue damage in both materials and scanning electron microscopy was used to examine the specimens before and after testing. For static testing, the resonators were subjected to external loading, which was then fixed using adhesive.
Experiments with silicon films investigated two aspects: First, to confirm whether surface oxidation is the critical parameter in fatigue and in an attempt to engineer fatigue resistant silicon films, the influence of nano-scale Al2O3 diff usion barrier coatings on cyclic fatigue was investigated. The Al2O3 coating lead to an increase in life by at least two orders of magnitude compared to uncoated devices in the harsh environment, which reaffirms reaction layer fatigue (RLF) as underlying model. Second, previous low cycle fatigue (LCF) data was inconsistent with the RLF model, given that thick surface oxidation is unrealistic for tests lasting only few minutes. Accordingly, static fatigue in silicon was investigated as underlying cause. Results indeed suggest that time-dependent crack growth is responsible for LCF failures (< 1 x 10^7 cycles, corresponding to about 5 min at 40 kHz).
The influence of environment on nickel has previously been studied, but not specifically for lifes in the very high cycle fatigue regime (VHCF, > 1 x 10^8 cycles), which corresponds to typical lifetimes of MEMS. Consequently, this aspect has been investigated in the present study. Preliminary results suggest that oxidation may indeed factor into the VHCF behavior of nickel.