Fabrication of free-standing MEMS is fraught with obstacles. Residual stresses which present themselves during various phases of the depositions, and etching processes can have severe effects on the performance of MEMS devices. Internal mechanical stresses cause the free-standing microstructures to warp, buckle up, bent down and ultimately touch down onto the supporting substrate, thus deteriorating the performance of the microsystem or even causing complete failure (Cianci, Coppa, and Foglietti, 2007). In addition, the multilayer nature of most MEMS devices causes additional residual stress to present itself because of the variations in material properties between two layers; the inconsistencies in thermal expansion between layers exacerbates the residual stress issues. Attempts to control residual stress using different fabrication techniques are being implemented, and have been successful in reducing the negative effects of stress. However, development of optical microstructures such as microbolometers needs nanometer scale control of thin-film shape, thus greater control of residual stresses (Bifano, Johnson, Bierden, and Mali, 2002). Hence, a suitably accurate and efficient method for residual stress measurement is required.
The Sensors and Integrated Microsystems lab (SIMSLab) is currently focusing on a method for the reliable characterization of residual stress caused by fabrication. In-situ stress measurement devices (Fig. 1) are being used to extract accurate values of residual stress. Since MEMS are fabricated using a wafer deposition process, the use of in-situ devices allows us to gather localized measurements of stress over the entire surface of the wafer (Fig. 2 and 3). Knowing the local residual stress trends in a wafer, we are able to identify wafer sites which minimize residual stress effects, allowing designers to increase the fabrication yield of their devices.
