Fast Scanning and Fast Image Reconstruction in Atomic Force Microscopy

摘要

This project aims at addressing challenges of high throughput applications of microcantilever based devices which are predicted to have enormous impact on science, technology, and applications that include various defense applications. This project addresses the challenges in two steps - (1) to substantially increase the imaging throughput by closely packing an array of microcantilevers for parallel (2) to develop numerical algorithms to construct high resolution images from the scan data (typically corrupted with noise, blurring effects and tip-sample convolutions) that are time and storage memory efficient. With respect to (1), we have developed model microcantilever arrays with strong crosscoupling, compared control design decentralized and distributed controller architectures, presented new theory to analyze stability and performance of spatially and temporally varying plants, studied invariant nominal models and present necessary and sufficient conditions for robustness when the underlying perturbations on them are spatiotemporal varying linear or nonlinear and unstructured or structured. We addressed performance and robustness issues with noisy and failed communications between subcontrollers and issues with performance under switching and reconfiguring. With respect to (2), we developed robust numerical algorithms that solve integral equations defined on irregularly (as opposed to rectangular) shaped domains which model the scan data from most AFM application.