Keywords: MEMS self-calibration, simulated experiment
Summary:This talk presents a new concept called a simulated experiment, where a MEMS simulation is treated as it would be treated in a real experiment. This is done by assuming the following: * Geometric overcuts and fillet sizes are unknown * Material properties are unknown * Asymmetries about cross sections are unknown * Residual and packaging stresses are unknown * Parasitic capacitance changes between measurements * Dielectric charging changes between measurements * Temperatures drift between measurements. The simulated experiment proceeds by conducting measurements that are only available in a realistic setting; in this, measurements of capacitance due to applied voltages by electrical probing. Using only capacitance and voltage, we demonstrate how mechanical measurements of geometry, displacement, electrostatic actuation force, stiffness, mass, damping, and material properties may be obtained. This is followed by a determining the accuracy of such measurements. Accuracy is a measure of the difference between measurement and the true value. A unique benefit of a simulated experiment, which is currently difficult to do in real experiments of MEMS, is the determination of measurement accuracy. That is, although real experiments are able to produce an average value and a standard deviation from several measurements, since the true value is usually unknown in real experiments of MEMS, then accuracy is difficult to determine. The reason why true values are usually unknown in real experiments is due to the lack of measurement standards at the microscale or the lack of measurement methods that are traceable back to a macroscale NIST or ASTM standard. The rationale for studying MEMS self-calibration through a simulated experiment is to explore the accuracy of performing the method of electro micro metrology (EMM) on real experiments. We have previously compared the EMM measurement of geometric overcut in a MEMS device against measurement by scanning electron microscopy (SEM). The comparison showed that EMM resulted in a lower uncertainty than SEM. This presentation examines the accuracy of our EMM method and compares its accuracy, precision, and comprehensiveness to other measurement methods.