A Novel Experimental Technique for Testing Thin Films and MEMS Materials

We have developed a novel μ–scale membrane deflection experiment particularly suited for the investigation of submicron thin films and MEMS materials. The experiment cons ists of loading a fixed-fixed membrane with a line load that is applied to the middle of the span with a nanoindenter column. A Mirau microscope-interferometer is positioned below the membrane to observe its response to loading. This is accomplished through a specially micromachined wafer containing a window to expose the bottom surface of the membrane. The sample stage incorporates the interferometer to allow continuous monitoring of the membrane deflection during both loading and unloading. As the nanoindenter engages and deflects the sample downward, fringes are formed due to the motion of the bottom surface of the membrane and are acquired through a CCD camera. Digital monochromatic images are obtained and stored at periodic intervals of time to map the strain field. Through this method, loads and strains are measured directly and independently without the need for mathematical assumptions to obtain the necessary parameters for describing material response. Additionally, no restrictions on the material behavior are imposed in the derivation of the model. In fact, inelastic mechanisms including strain gradient plasticity effects can be characterized by this technique. INTRODUCTION Thin Films & MEMS: Thin films, thickness of a few microns or less, are applied as components in almost all MEMS devices and frequently serve as essential device functions. The demands placed on thin films in these applications can sometimes subject them to various mechanical conditions, such as; fracture, plasticity, friction and wear, creep, fatigue, etc.. Most knowledge of bulk material behavior fails to describe material response in this size regime. Many researchers currently have programs to study such characteristics [1-4]. Frequently, each particular investigation involving MEMS tends to be device dependent and introduces new fundamental questions. Progress in this field has leaned toward providing more specific technological solutions rather than generating a basic understanding of mechanical behavior. Testing Methodologies: Techniques to study MEMS materials response to mechanical loading are diverse and can be classified by as static or dynamic. Although both will yield the materials mechanical properties, they accomplish it in completely different manners. Within the static group are nanoindentation (in standard DC mode) [5], micro-tensile [6], bending [6-9] and bulge tests [10-12]. Nanoindentation (when the Continuous Stiffness Measurement is used), resonance and fatigue methods [13-15] belong to the dynamic group in this observation. Conventional understanding of yielding does not apply at this scale because of the increased role that interface driven processes play. Thus, there is a need to establish novel testing methodologies that contain no mathematical assumptions and measure material parameters directly and independently. The equivalent of a tensile test performed on bulk samples is desirable for thin films for several reasons. Loads and strains are measured directly and independently; no mathematical assumptions are needed to obtain the parameters describing the material response. Techniques that use a special fixture to load small tensile samples have been developed, [7-9]. However, stress-strain curves cannot be uniquely determined when the various techniques are compared. This is due in part because complex sensors and actuators for loading are used for data acquisition. An ideal architecture to achieve a direct tensile testing scheme involves a freestanding membrane that is fixed at both ends. A line load applied at the middle of the span would produce a uniform stretch on the two halves of the thin membrane. We have demonstrated this testing scheme by the investigation of RF (radio frequency) MEMS switches, produced by Raytheon Systems Co.[16-18]. In this method we made use of a nanoindenter to apply a line load at the center of the membrane. Pushing the membrane down tests the specimen structural response and provides information on its elastic behavior and residual stress state. In this manner, simple tension of the membrane is achieved except for boundary bending effects. The critical concern in this Membrane Deflection Experiment (MDE) was accounting for the thermal drift and spring constant of the nanoindenter column. Since the column dimension is orders of magnitude larger then the membrane 446 Proceedings of the SEM Annual Conference on Experimental and Applied Mechanics, June 4-6, 2001, Portland, Oregon