Study of repair of stiction failed micro cantilever beams using structural vibrations

Date of Completion

January 2007


Engineering, Mechanical




The large scale use of micro-electro-mechanical systems (MEMS) in commercial applications has been hampered by reliability issues (i.e., mechanical failures), which arise during the fabrication process and/or in field operations. The most prevalent of these failure mechanisms is adhesion, as neighboring components come into sticking contact. This is commonly stiction failure. In the past, efforts have been made to minimize stiction failures (i) during the fabrication process, by using materials with low adhesion energy or by applying hydrogen treatments, etc., and (ii) during operation (post-fabrication) using a laser release process. These methods have achieved some measure of success. However, both methods are expensive and make MEMS a less attractive option in the marketplace. As a result, there is a compelling need to develop an effective, reliable, and low cost method for achieving stiction repair. The focus of this thesis is to develop a new approach to this problem. This is a radical departure from previous methods and involves using mechanical vibrations to affect the desired repair. ^ The current study investigates the potential for using vibrations as a stick release mechanism; this work merges the areas of dynamics and fracture mechanics to study this method. There are three goals to this study. The first is to understand how mechanical excitation might be used to drive stiction failed micro-cantilevers and initiate the desired repair. Preliminary experiments, which parallel the analysis, provide evidence that this approach is viable. The second is to determine whether electrical excitation could be used to achieve similar success in initiating the repair of the failed cantilevers. The motivation behind this stems from the electrical capabilities of the MEMS chip itself, i.e. the functionality of the chip may be used to repair itself. The analysis that follows shows that the electrically repair process also works but that it is driven by a fundamentally different mechanism than that for mechanical loading. Lastly, this work focuses on what happens after the repair is initiated. Specifically, it looks at how the debonding process proceeds dynamically. The study shows scenarios for partial or complete repair. The implications for all of these behaviors are discussed thoroughly in terms of the underlying mechanics of the system. ^