Design, sensing and control of monolithic nanopositioning devices

The rapidly growing field of nanotechnology has increased the need for high-speed nanopositioning systems with sub-nanometer resolution. These devices are critical in applications including scanning probe microscopes, precision optics, and lithography. A typical nanopositioner is comprised of three elements: a piezo stage, a position sensor, and a controller. To achieve nanoscale resolution with high accuracy, each of these elements needs to be carefully designed and optimized. This study introduces a new class of nanopositioning devices that are constructed from a single or laminated sheet of piezoelectric material. Active flexures that create actuation force and guide the motion are created by subtractive machining and electrode patterning. Compared to flexure based devices, the proposed approach has a much lower vertical size and weight, and is significantly easier to fabricate. This approach is used to design nanopositioners with two, three, and six degrees of freedom. The behavior of these devices is assessed by analytical modeling, finite element analysis, and experimental characterization. The performance of a nanopositioning system is also restricted by the characteristics of position sensors. Among the various types of sensing techniques available in the literature, piezoelectric and piezoresistive sensors are best suited for integration into low-profile nanopositioning devices. For piezoelectric sensors, a new sensing technique is demonstrated which compensates for the high-pass response, allowing the direct application of integral feedback control. For piezoresistive sensors, a novel sensing technique is proposed which allows independent estimation of temperature and strain in tee-rosette piezoresistive strain sensors. Furthermore, a novel sensing method for five degrees-of-freedom (DOF) displacement measurement using piezoresistive sensors is introduced. The proposed technique employs a combination of feedforward and system identification techniques to estimate linear and angular displacements in five DOFs. Due to the excitation of mechanical resonances, tracking of reference trajectories has been the main challenge in the control of piezoelectric nanopositioners. Inversion-based feedforward techniques have been widely used in the literature to invert system dynamics. The foremost difficulties associated with plant inversions are model uncertainties and non-minimum phase zeros. To overcome these difficulties a model-less approach using FIR filters is presented. The results demonstrate the effectiveness of the feedforward technique in reducing cross-coupling and achieving significantly improved output tracking.