A Non-raster Scanning Approach in Atomic Force Microscopy Using a Combined Contour Prediction Algorithm

In this paper, we present a novel non-raster scanning algorithm for high-speed imaging in Atomic Force Microscopy. In contrast to recent non-raster scanning algorithms for string-like samples, the proposed algorithm is developed for cells and other simple specimen samples. This algorithm collects data in the vicinity of the specimen to create sample contours at different heights to build the 3D topography of the target sample. During the scan process, the tip is steered based on a prediction of the contour curvature and contour tangent. The proposed scanning trajectory follows the contour of the sample and avoids
crossing the specimen, while minimizing the possible excitation of resonances of the cantilever. For the prediction of the curvature and tangent of the contour, the current partially obtained contour and a previous contour scan are used: a prediction from both contours is suitably combined by a weighting algorithm
derived from a reliability evaluation of both predictions. This permits the creation of topographical images of specific interest at a reduced scanning time in comparison to some prevalent non-raster scan algorithms and raster scans. Simulation results are provided.
